High Molecular Weight (HMW) Dissolved Organic Matter (DOM)
In Seawater: Chemical Structure, Sources and Cycling
by
Lihini I. Aluwihare
B.A., Mount Holyoke College
(1993)
SUBMITTED IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR
OF PHILOSOPHY
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
and the
WOODS HOLE OCEANOGRAPHIC INSTITUTION
April, 1999
© Lihini I. Aluwihare, 1999. All rights reserved.
The author hereby grants to MIT and WHOI the permission to reproduce and to distribute copies of this
thesis document in whole or in part.
I2 ,J
/I
A
Signature of Author:
ciences,
egrric an
Department of Earth, Atmo
Massachusetts Institute of Teehn oT6gy and the Joint Program
in Oceanography, Massachusetts Institute of Technology/Woods
Hole Oceanographic Institution, April 30, 1999.
Certified by
-
Daniel J. Repet'a
Thesis Supervisor
Accepted by
,
,,
Edward A. Boyle
Chairman, Joint Committee for Chemical Oceanography,
Massachusetts Institute of Technology/Woods Hole
Oceanographic Institution
LIBRARIES
ACKNOWLEDGEMENTS
I would like to thank my advisor, Dan Repeta, for his extraordinary commitment
to this thesis. I have benefited greatly from his high expectations and his high scientific
standards. I would also like to thank my thesis committee, Phil Gschwend, Charles
Hopkinson and Cindy Lee who have all made invaluable contributions to this work. Tim
Eglinton and Jim Moffet have both been instrumental to my survival in the joint program,
offering fresh perspectives and a welcome sense of humour. I would also like to thank
Tim for being a wonderful role model. Thanks also to Carl Johnson, Big "Bob" Nelson,
Nelson Frew, Ollie Zafiriou, Jeff Seewald and Lorraine Eglinton, all of whom have left
their marks on this thesis. I would like to thank the members of Fye Lab in general, for
their sense of humour and compassion throughout my six eventful years here.
Many thanks to my friends here in Woods Hole: Kathy Barbeau, Liz Koodge,
Mak D. S. Saito, Lil' Ann Pearson, the "24 Millfield bunch" (Danny Sigman, Garret Ito,
Ellen Ito, Jon Woodruff, Joe Warren, Monica Relle, Kira Lawrence, Peter Sauer, Kitty,
Zorro and Patches), Kirsten Laarkamp, Dana Stuart, CHRIS REDDY, Silvio Pantoja,
Steve Jayne, Jess Adkins and Mike Atkins, for all the good times, support and
encouragement.
Finally to my extended family: I made it because of you. I cannot find the words
to thank my beloved parents who invested everything in me: ammi, I am everything I am
because you loved me; apuchi, your incredible insight and dedication to education
inspired me to reach this goal. To ayia, thank you for being my guardian angel. To
Phyllis and Peter, your love and quiet support have carried me throughout the last six
years. To my family in Sri-Lanka, I am lucky to be so loved. To Bill Shaw, our time
together has been paradise, thank you for all your patience and support. To Emily Norton,
the closest thing I have to a sister, Dianne Keller, Monica Chander and Laura Termini,
thanks for still putting up with me after ten years. To the Arvons, the family of strong
women, especially little Dr. Regina, thank your for your inspiration.
This thesis is dedicated to all the mothers who inspired their daughters to move
mountains.
This thesis was funded by a grant from the US Department of Energy, Ocean
Margins Program.
High Molecular Weight (HMW) Dissolved Organic Matter (DOM) in
Seawater: Chemical Structure, Sources and Cycling.
Lihini I. Aluwihare
Abstract
The goal of this thesis was to use high resolution analytical techniques coupled
with molecular level analyses to chemically characterize high molecular weight (> 1 k Da
(HMW)) dissolved organic matter (DOM) isolated from seawater in an attempt to provide
new insights in to the cycling of DOM in the ocean.
While a variety of sites spanning different environments (fluvial, coastal and
oceanic) and ocean basins were examined, the chemical structure of the isolated HMW
DOM varied little at both the polymer and monomer levels. All samples show similar
ratios of carbohydrate:acetate:lipid carbon (80±4:10±2:9±4) indicating that these
biochemicals are present within a family of related polymers. The carbohydrate fraction
shows a characteristic distribution of seven major neutral monosaccharides: rhamnose,
fucose, arabinose, xylose, mannose, glucose and galactose; and additionally contains Nacetylated amino sugars as seen by Nuclear Magnetic Resonance Spectroscopy (NMR).
This family of compounds, consisting of a specifically linked polysaccharide backbone
that is acylated at several positions, has been termed acylated polysaccharides (APS) by
our laboratory. APS accounts for 50% of the carbon in HMW DOM isolated from the
surface ocean and 20% of the carbon in HMW DOM isolated from the deep ocean.
In order to identify a possible source for APS three species of phytoplankton,
Thalassiossiraweissflogii, Emilianiahuxleyi and Phaeocystis,were cultured in seawater
and their HMW DOM exudates examined by variety of analytical techniques. Both the T.
weissflogii and E. huxleyi exudates contain compounds that resemble APS indicating that
phytoplankton are indeed a source of APS to the marine environment. Furthermore, the
degradation of the T. weissflogii exudate by a natural assemblage of microorganisms
indicates that the component resembling APS is more resistant to microbial degradation
compared to other polysaccharides present in the culture.
Molecular level analyses show the distribution of monosaccharides to be
conservative in surface and deep waters suggesting that APS is present throughout the
water column. In order to determine the mechanism by which APS is delivered to the
deep ocean the A14 C value of APS in the deep ocean was compared to the A14C value of
the dissolved inorganic carbon (DIC) at the same depth. If the formation of deep water is
the dominant mode of transport then both the DIC and APS will have similar A' 4C
values. However, if APS is injected into the deep ocean from particles or marine snow
then the A14 C value of APS will be higher than the DIC at the same depth. Our results
indicate that APS in the deep Pacific Ocean carries a modem A 4 C value and is
substantially enriched in 14C relative to the total HMW DOM and the DIC at that depth.
Thus, particle dissolution appears to be the most important pathway for the delivery of
APS to the deep ocean.
Thesis Supervisor: Daniel J. Repeta
TABLE OF CONTENTS
TABLE OF FIGURES .................................................................................................................................
7
11
TABLE OF TABLES ....................................................................................
1.
INTRODUCTION ..............................................................................................................................
1.1
1.2
1.3
1.5
1.5
1.6
1.7
13
.................... 13
THE ROLE OF DOM IN THE CARBON CYCLE .........................................................
......................... 14
SOURCES OF DO M .......................................................................................
........... 16
...........................
...............
CYCLE OF DOM IN THE OCEAN
19
......................
CHEMICAL STRUCTURE OF DOM ........................................
.......... 20
..........................
................
ORGANIZATION OF THIS THESIS
.. .......... .............................. 23
GOALS OF THIS THESIS ......................................
REFERENCES....................................................................................... 24
STRUCTURAL CHARACTERIZATION OF MARINE DISSOLVED HIGH MOLECULAR
2.
31
WEIGHT ORGANIC MATTER ..............................................................................................................
31
................................
INTRODUCTION ..................................
2.1
M ATERIALS AND M ETHODS........................................................................................................... 33
2.2
33
Sampling Site ............................................................................
2. 2.1
33
Sample Collection and Isolation................................................
2.2.2
...... ............... 35
U ltrafiltration .........................................................
2.2.2.1
............ ......... 38
Cleaning The Membranes....................................
2.2.2.1.1
Diafiltration....................................................................................... 38
2.2.2.2
.......... ............ 39
Desalting .......................................................
2.2.2.3
Dissolved OrganicCarbon (DOC)Analysis................................................ 40
2.2.3
..... .. 40
Elem ental Analysis.........................................................................................................
2.2.4
.................... 41
Nuclear Magnetic Resonance Spectroscopy........................
2.2.5
... 41
...............................................
General Principles.................................... ............
2.2.5.1
42
........................
Experimental Conditions for Proton NMR ('HNMR).........................
2.2.5.2
3
....................... 44
C and NN M R ...........................................................................................
2.2.5.3
45
......................
...............
CarbohydrateAnalysis ..........
2.2.6
M onosaccharide Analysis .............................................................. ........... ............. 45
2.2.6.1
................... 46
L inkage Analysis ....................................................................................................
2.2.6.2
49
LipidAnalysis ...............................................................
2.2. 7
.................... 50
..................................
Am ino Acid Analysis .................. ..............
2.2.8
....... ................................ 51
RESULTS ............................................
2.3
A ncillary D ata ...................................................................................................................... 5 1
2.3.1
53
N uclear Magnetic Resonance Data......................................................................................
2.3.2
............ ................... 53
Proton NM R........................................
2.3.2.1
63
N itrogen NM R .........................................................................
2.3.2.3
67
..................
......................................
..........
Lipid Results ....................
2.3.3
69
MonosaccharideAnalyses .............................................
2.3.4
.. 77
..........................
M ethylation Analysis............................................................
2.3.5
............ 80
Linkage Analysis ...................................
2.3.6
.................. 87
Am ino Acid Analysis .......................................................................................
2.3.7
...................... 89
D ISCUSSION ............................................
2.4
.................. 96
CONCLUSIONS .............................................
2.5
REFERENCES........................................................................ 98
2.6
105
........................
A PPEN DIX ........................................................................................................
2.7
105
..........
...
Conditions............................
Spectroscopy
Resonance
Magnetic
Nuclear
2.7.1
106
.............................
Linkage Datafor Seawater.................... ............
2.7.2
...... ..................... 108
Retention Times for PartiallyMethylatedAlditol Acetates... ... ...
2. 7.3
3. A COMPARISON OF THE CHEMICAL CHARACTERISTICS OF OCEANIC HMW DOM
111
AND EXTRACELLULAR DOM PRODUCED BY MARINE ALGAE. .........................................
3.1
3.2
IN TRO DUCTION ............................................................................................................................
111
M ATERIALS AND M ETHODS............................................................................
........................ 113
...................................
Phytoplankton Cultures .... .......
Ultrafiltration.............................................................
A n alyses ..............................................................................................................................
Dissolved Oxygen Measurements.....................................................................................
3.2.1
3.2.2
3 .2.3
3.2.4
3.3
RESULTS ...........................................
3.3.1
3.3.1.1
3.3.1.2
3.3.1.3
.......... 118
118
...................................
POC, DOC, and HMW DOC .........................................
'H N M R D ata.....................................................................................................
M onosaccharide Data ..............................................
BacterialDegradationof the T weissflogii Exudate ...............................
3.3.2
3.4
............
Phytoplankton Cultures ...........
............ 118
....................
122
124
.......
..................
C hanges in DO C .................................................................................................
3.3.2.1
..............
'H N M R D ata...........................................................................................................
3.3.2.2
M onosaccharide Data.....................................................................................................................
3.3.2.3
3.3.2.4
Am ino A cid D ata ...........................................................................................................................
DISCUSSION .................................................
.......... ..............
126
126
128
130
132
134
................ 134
143
...........................
Phytoplankton Cultures......
Degradationof the T weissflogii exudate ......................................
3.4.1
3.4.2
113
117
1 17
117
CONCLUSIONS ......................................................... ....
.................. 150
REFERENCES..
..................................................
.............
......................
152
A PPEN D IX ....................................................................................................... 160
3.5
3.6
3.7
3.7.1 Linkage datefor the Cultures......
160
.......................................
4. RADIOCARBON VALUES OF INDIVIDUAL MONOSACCHARIDES ISOLATED FROM
163
OCEANIC HMW DOM .................................................................................................................
4.1
4.2
INTRODUCTION ....................................
.................................
................... 163
MATERIALS AND METHODS............................................................ 166
......... ........... 166
Sampling ......................................
168
..........................................
ChemicalAnalyses
High-PerformanceLiquid Chromatography (HPLC)................................... 170
4.2.1
4.2.2
4.2.3
4.2.3.1
4.2.3.2
4.2.4
4.3
RadiocarbonAnalyses ..............................
DISCUSSION .................................................
4. 4.1
4. 4.2
4.5
4.6
................... 170
173
Sample Preparationfor AcceleratorMass Spectrometry (AMS) ...................................... 177
RESU LTS ...................................................................................................................
4.3.1
4.4
.........
Ion Exchange .........................................
Normal Phase Chromatography ........................................
...
................................
..........
..............
................. 195
Radiocarbon Results.................................................... ......... ........... 195
201
Analytical Methods ...............................................................................
.................
CONCLUSION S ....................................................................................................
REFERENCES.......................................................................
CONCLUSIONS ...........................................................................................................................
5.
5.1
5.2
5.3
18 1
.................. 191
204
206
211
G ENERAL CONCLUSIONS .................................................................................. ...................... 211
219
FUTURE RESEARCH DIRECTIONS .......................................................
REFERENCES FOR CHAPTER 5 ................................................................................................... .. 223
TABLE OF FIGURES
Chapter 1
Figure 1.1 Depth profiles of dissolved organic carbon (DOC) in the Sargasso Sea (filled diamonds) and the
...... 15
North Pacific Ocean (open squares) (data from Druffel et al. 1992) .....................
Figure 1.2 The A'4 C values of DOC and DIC (in parentheses) in different parts of the Worlds Ocean.
......................
Arrows depict the direction of ocean circulation ..............................................................
18
Figure 1.3 Hypothetical structure for dissolved marine humic substances (Gagosian and Stuermer, 1977) 19
Figure 1.4 Monosaccharide structure: carbon and protons are labeled 1-6 as shown (a), 13anomer (b) and a
21
............................................................................
anomer (c)
Figure 1.5 Eight neutral monosaccharides identified in HMW DOM: hexoses (A); deoxy hexoses (B);
pentoses (C). Glucosamine in its acetylated and non-acetylated form (D) is also shown. ....................... 22
Chapter 2
Figure 2.1 Map showing the locations of the different samples analyzed for this thesis........................... 34
Figure 2.2 Schematic representation of the procedure for the isolation of HMW DOM .......................... 36
....... 46
Figure 2.3 Synthesis of alditol acetates from monosaccharides .......................
Figure 2.4 The reaction scheme for the preparation of partially methylated alditol acetates for linkage
analysis of polysaccharides.......................................................................................................
48
Figure 2.5 The major 'HNMR resonances identified in HMW DOM. The baseline for each of the major
54
peaks and the whole spectrum are shown on the figure as dashed lines ..........................................
Figure 2.6 Two 'HNMR spectra for surface HMW DOM isolated from the North Atlantic (MAB943) (A)
and North Pacific (B) (P961). Note the similarities in the resonances present and the relative proportion of
56
................................................
.....................................................
each resonance .
Figure 2.7 The figure shows 'HNMR spectra for a typical surface ocean sample (A) (CRL), a typical deep
ocean sample (B) (P962) and a spectrum seen for two of the eight deep samples analyzed (C) (MAB944).
Dashed lines in Spectrum B show the baseline for each of the identified peaks ...................................... 57
Figure 2.8 'HNMR spectra of MAB948, a HMW DOM sample from a depth of 750 m before desalting
59
(A), after desalting with a cation exchange resin (B) and after hydrolysis (C)...................................
Figure 2.9 Proton NMR spectrum of a 5:1 mixture of maltoheptose and Gramicidin S, in D20. Resonances
enclosed in brackets arise from proteins.............................................................. 60
Figure 2.10
13CNMR
spectrum of HMW DOM in D20 ........................
.
......
..........
62
Figure 2.11 Solid state 15NNMR spectra of HMW DOM , before (A) and after (B) 1 M HCI hydrolysis for
11 h at 105 0 C. Resonances are reported relative to NH 3. Pyridine was added as an external standard to
............... 64
quantify the total nitrogen detected by 'SN NMR. ..................................................................
0
Figure 2.12 'H NMR spectra of HMW DOM, before (A) 1 N HCI hydrolysis for 11 h at 105 C and after
.............. 65
hydrolysis and freeze drying to remove acetic acid (B)......................................
Figure 2.13 'H NMR of HMW DOM prior to 4 N HCI hydrolysis (A), following hydrolysis (B), the
aqueous layer (pH = 3) after extraction with ethyl ether (C) and the ethyl ether layer (D) ....................... 68
Figure 2.14 Raw gas chromatograms of alditol acetates. ....................
........................
70
Figure 2.15 The mole percent distribution of the seven neutral monosaccharides identified in HMW DOM.
...................
72
Figure 2.16 Correlation coefficient data for the mole percent monosaccharide distribution of each HMW
..................... 75
DOM sample. All correlations are reported relative to MAB941 .............................
Figure 2.17 A schematic representation of the flow of water in the MAB constructed using the correlation
76
.......................
coefficient data for the mole percent monosaccharide distributions......................
Figure 2.18 The 'H NMR spectrum of a standard mixture of monosaccharides (A) and a hydrolyzed HMW
........ ................. 78
DOM sam ple (B). ........................................................
Figure 2.19 'HNMR spectra of HMW DOM prior to methylation (A) and following methylation and
.................. ............. 79
extraction into an organic solvent (B)...................................
Figure 2.20: The total ion current of partially methylated alditol acetates from a HMW DOM sample. Each
peak was identified using mass spectral data and retention time data .................................... 81
Figure 2.21 Mass spectra of terminally linked rhamnose (A), terminally linked mannose (B), 1,4-linked
galactose (C ).........................................................................................................................................
84
Figure 2.22 The relative distribution of different types of linkages per sugar for three different HMW DOM
....... ................. 86
samples. ..........................................................
Figure 2.23 The total amount of carbon in surface water HMW DOM derived from refractory deep ocean
HMW DOM, and HMW DOM newly added at the surface (A) and the relative amount of carbon accounted
for by the hydrolysis products (B), assumed to be derived from new HMW DOM only .......................... 91
Figure 2.24 The total carbon in surface water HMW DOM accounted for by the biochemicals identified in
95
.........................
this study. Also refer to Figure 2.23 ........................................
Figure 2.25 Hypothetical structure for APS, containing neutral monosaccharides and N-acetylated sugars.
95
...................
Chapter 3
Figure 3.1 Experimental set up for the culture experiments. Seawater was pre-filtered through 0.2 jim and
1 nm membranes. Both the Emilianiahuxleyi and Phaeocystis experiments were terminated after the algae
were removed without further degradation of the exudate. TW3, TW4, TW5 and TW6 refer to samples
taken during the degradation of the ThalassiosiraWeissflogii exudate (dates are given in Figure 3.3)..... 115
Figure 3.2 Growth curve for the E. huxleyi culture expressed in cell numbers/ml...........................
118
Figure 3.3 Ancillary data from the Thalassiosiraweissflogii culture. (A) Changes in chl-a (filled squares)
and POC (filled triangles). (B) Changes in the DOC (filled squares) concentration and C/N ratios of the
POC (open triangles). Algae were removed on day 8. TW1, refers to the ultrafiltration sample taken during
early log phase, TW2 was taken during late log phase, and TW3, 4, 5 and 6 were taken during the course of
the degradation experiment after the algae had been removed. After day 9, changes in DOC concentration
result from the action of microorganisms present in the incubation, or from physical removal mechanisms.
121
...................
Figure 3. 4 'HNMR spectra for HMW DOM isolated from seawater (Woods Hole) (A), and the exudates
of Thalassiosiraweissflogii (B), Emiliania huxleyi (C), and Phaeocystis (D). Carbohydrate resonances are
between 3.2- 4.7 p.p.m (CHOH), 5-5.8 p.p.m (anomerics) and 1.3 p.p.m (CH 3 from deoxy sugars). Spectra
also show proton resonances for acetate (2.0 p.p.m) and lipids (1.3 and 0.9 p.p.m). The ratios of
carbohydrate:acetate:lipid reported in the text are based on the areas under the peaks denoted C (for
carbohydrates), Ac (for acetate) and Alkyl (lipids). To convert from hydrogen to carbon we use a C/H ratio
of 1 for carbohydrates, 0.67 for acetate and 0.5 for lipids. ................................................ 123
Figure 3.5 The relative distribution of monosaccharides in HMW DOM isolated from seawater, and the
exudates of Thalassiosira weissflogii, Emiliania huxleyi, and Phaeocystis. R is rhamnose, F is fucose, A is
arabinose, X is xylose, M is mannose, GL is glucose, GA is galactose, and RI is ribose. The seawater
sample represents the average distribution of monosaccharides in ten seawater samples (the standard
deviation is less than ±3% for each monosaccharide) .........................................
126
Figure 3.6 'HNMR spectra of two HMW DOM samples isolated during the degradation of the
Thalassiosiraweissflogii exudate, TW3 (A) and TW4 (B). Resonances between 1.9-3.1 p.p.m are thought
to arise from proteins which can be removed by passage through a cation exchange column (insets). ..... 129
Figure 3.7 The relative distribution of amino acids in HMW DOM isolated during the degradation of the
Thalassiosiraweissflogii exudate, expressed as the percentage of each amino acid accounted for by the
culture sample. Amino acid are: aspartic acid, glutamic acid, serine and histidine, glycine, threonine and
arganine, alanine, tyrosine, methionine, valine, phenylalanine, isoleucine, leucine and lysine............... 133
Figure 3.8 Carbohydrate linkage data for the exudates. TW2, E. huxleyi and Phaeocystissamples were
taken when the cultures were harvested. TW4 was isolated during the Thalassiosiraweissflogii incubation
experiment. The number beneath each bar refers to the linkage site on each monosaccharide (i.e. carbon 1
and 6 as 1,6), and T refers to linkages through carbon 1 only. Only linkages which constitute > 5% of each
........................... .............. 139
sugar are listed .
...................................................
Figure 3.9 The change in the concentration of total neutral monosaccharides in HMW DOM and the change
in APS as a percentage of the total neutral monosaccharides (as calculated in the text) over the course of the
............. 147
Thalassiosiraweissflogii incubation experiment.............................
Chapter 4
Figure 4.1 Two scenarios for the delivery of surface produced DOM to the deep ocean. Scenario A shows
the delivery of DOM to the deep ocean via deep water formation. Scenario B shows that in addition to
advection, DOM may also be delivered to the deep ocean via bacterial degradation and disaggregation of
......................... 165
sinking POM.....................................
Figure 4.2. A flow chart of the various analytical steps involved with the separation of hydrolyzed
monosaccharides from HMW DOM. Boxes in bold denote HPLC steps. Silver column refers to the cation
exchange steps and amino column refers to the reverse phase separation ........................................ 169
Figure 4.3 (A) Reproducibility of six consecutive injections of a monosaccharide mixture on the Ag
column. Standard deviations (% Stnd. Dev) for retention times (Ret. Times) and peak areas are reported as
a percentage. (B) Calibrated detector response for the three different monosaccharide fractions (R2 for each
linear fit to the data are reported on the figure). ....................................................... 173
Figure 4.4 (A) Reproducibility of six consecutive injections of a monosaccharide mixture on the amino
column. Standard deviations (% Stnd. Dev) for retention times (Ret. Times) and peak areas are reported as
a percentage. (B) Calibrated detector response for five monosaccharide fractions (R2 for each linear fit to
176
.........................
the data are reported on the figure) ........................................
Figure 4.5 CRL sample after hydrolysis and neutralization, separated on the cation exchange column
using Ag as the counterion. The figure shows the sample before (A) and after (B) desalting using an anion
...... .................. 182
exchange resin. ..........................................................
Figure 4.6 HPLC of CRL on the Ag column, during the initial separation (A) and the second separation
.................. 187
...... ..
. ........................................................................
(B )...................
Figure 4.7 Th separation of individual monosaccharides from unhydrolyzed HMW DOM using a
combination of cation exchange and reverse phase HPLC. (A) Separation of glucose and rhamnose; (B)
separation of mannose, xylose and galactose; (C) separation of fucose and arabinose .......................... 190
Figure 4.8 Radiocarbon values for different fractions of carbon in the North East Pacific (Station M).
POCsu, is the Al 4 C value of suspended POC, and POCsink is the Al4 C value of sinking POC. Depth profiles
show DOC and DIC A' 4C values; filled circles are the A14C value of total HMW DOC (surface and 1600
m), and filled diamonds are the A' 4C value of the HMW Sugars (surface and 1600 m) DIC, POC and total
196
DOC values were obtained from Druffel et al. 1996 and Bauer et al. 1998 ......................................
Figure 4.9 Variations in the concentration of total DOC ( tM) over the whole ocean (after Hansell and
................... 198
Carlson, 1998). ..................................................
Figure 4.10 A comparison of the distribution of monosaccharides in HMW DOM (Average), Transparent
Exopolymer Particles (Avg TEP) ( after Mopper et al., 1995), and sediment trap material (POC) (after
Cowie and Hedges, 1984). Monosaccharides are galactose (Ga), fucose (F), mannose (M), glucose (G),
rhamnose and arabinose (R+A) and xylose (X) ...................................................... 200
Figure 4.11 The relative distribution of monosaccharides in HMW DOM injected onto the amino column
compared to the relative distribution of monosaccharides in HMW DOM determined by GC (AA). ....... 204
Chapter 5
Figure 5.1 Cycling of APS in the Ocean. ..................................................................................
.......... 215
Figure 5.2 Th cycling of APS in the Surface Ocean. See text for an explanation of the terms. ............ 216
TABLE OF TABLES
Chapter 2
Table 2.1 Ancillary data for each of the samples analyzed as part of this thesis ...................................... 39
Table 2.2 Relative abundance (in carbon) of the major biochemicals identified in HMW DOM by 1HNMR.
43
...............................................
.....................................................................................
Table 2.3 The relative distribution of neutral monosaccharides (as analyzed by alditol acetates) in HMW
............. 58
DOM . ..........................................................
Table 2.4 Results from the correlation coefficient analysis of the mole percent monosaccharide
..................... 75
distributions in different HMW DOM samples ....................
Table 2.5 The mole percent distribution of amino acids in HMW DOM ......................................
100
Chapter 3
Table 3.1 Ancillary data for the HMW DOM isolated from the phytoplankton cultures
........................ 106
Table 3.2 The relative distribution of monosaccharides in T weissflogii, E. huxleyi, and Phaeocystis at the
................... ......... .............. 112
tim e of harvesting . .......................................................
Table 3.3 Absolute concentrations of monosaccharides at the time of harvesting (TW2) and at each of the
time points during the incubation experiment (TW3, TW4, TW5 and TW6) for Thalassiosira
weissflogii. % C refers to the percent of total carbon in the sample accounted forby each of the
m onoosaccharides. ............................................................. ...... .................. 131
Table 3.4 Concentrations of the individual amino acids in HMW DOM isolated during the degradation of
the T weissflogii exudate (TW3, TW4, TW5 and TW6). .......................................................... 132
Table 3.5 Monosaccharide composition of polysaccharides isolated from different species of algae........ 142
Table 3.6 Degradation rate constanst for the different organic carbon fractions in the T.weissflogii over the
148
course of the incubation experiment.......................................................
Chapter 4
Table 4.1 Sampling site and DOC data for the samples analyzed in this study ...................................... 167
Table 4.2 The recovery of carbon from the different purification steps performed in this analysis. Desalting
was performed on an anion exchange column. Ag column and Amino column data, refer to
183
recoveries during the HPLC procedure ........................................
Table 4.3 The A'4C values of the different carbon fractions for each of the sites examined in this chapter.
14
Refer to the text for descriptions of fm, fm* and A C.................................................. 192
Chapter 5
Table 5.1 Numerical values for each of the terms shown in Figure 5.2 ................
................... 208
12
1.
Introduction
1.1
The Role of DOM in the Carbon Cycle
Dissolved organic matter (DOM) constitutes the largest pool of reduced organic
carbon in the ocean (Hedges, 1992) and has long been recognized to be an important part
of the global carbon cycle. DOM is defined as the fraction of total organic matter in the
ocean that passes through a 0.2 pm filter. Seven hundred Gt of carbon are present in
seawater as total organic carbon. Of this, 90% exists as DOM (Hedges 1992). This
reservoir is nearly equal in size to the amount of organic carbon in the terrestrial
biosphere, and the amount of carbon in the atmosphere as CO 2.
DOM has been shown to play an important role in a variety of oceanic processes.
For example, DOM has been shown to be important in supporting microbial production.
Between 15-25 Gt C/yr move via DOM into bacteria and the "microbial loop" (Azam et
al., 1983; Fuhrman and Azam, 1982; Hagstrom et al., 1979). These large flux estimates
are based on the high rates of heterotrophic bacterioplankton production throughout the
surface ocean, and the high turnover rates observed for simple molecules such as amino
acids and monosaccharides in seawater (Bada and Lee, 1977; Kirchman et al., 1991;
Smith et al., 1992).
Apart from the importance of DOM as a substrate for bacteria, it has also been
shown that DOM may be directly utilized by phytoplankton as a nutrient source. Recent
work has shown that many phytoplankton have the capability to utilize DON for growth
either by cell-surface deamination, amino acid transport systems or proteolysis (Martinez
and Azam, 1993; Palenik and Morel, 1990). Studies have shown that the "brown tide"
causing unicellular algae (Aureococcus anophagefferens) can use organic nitrogen
substrates in the field for growth, thereby giving this organism a possible ecological
advantage (LaRoche et al., 1997).
DOM has also been shown to be important in determining the distribution of trace
metals in the ocean. During the last several decades, convincing evidence has accrued
demonstrating that a number ofbioactive trace metals are strongly influenced by organic
complexation. For example, >90% of the iron, copper and zinc in oceanic surface waters
exists chelated with organic ligands (Bruland, 1989; Coale and Bruland, 1988; Moffet
and Zika, 1987). It has been further suggested that the mean concentration of iron below
500 m in the ocean is primarily controlled by the complexation of iron with organic
ligands (Johnson et al., 1997).
DOM has further been shown to play a role in the reduction of metals in the
surface ocean. Organic complexes of Fe (III) can be reduced to Fe (II) by ligand to metal
charge transfer reactions resulting from the absorption of both visible and ultraviolet light
by DOM (Voelker et al., 1997). This absorption and fluorescence of chromophoric (or
colored) DOM (CDOM) is particularly important because it can affect ocean color and
thus interferes with the interpretation of remote sensing data (Blough et al., 1993).
1.2
Sources of DOM
Despite the importance of marine DOM, little is known about its structure and its
cycling. Depth profiles of DOC concentration in the ocean (Figure 1.1) have a
characteristic shape. In surface water concentrations are between 60-80 gM. Below the
euphotic zone, concentrations decline to between 35-45 gM by 600-1000 m, and remain
relatively constant throughout the deep ocean (Peltzer and Hayward, 1996). The
increased concentrations of DOM in surface waters indicates a surface source, and likely
results from a decoupling of production and removal processes over annual to decadal
time scales (Carlson and Ducklow, 1995). Although the mechanisms by which DOM is
added to surface waters is not well known, it has long been recognized that this carbon
represents by-products generated by algal photosynthesis in the euphotic zone (Duursma,
1963; Mague et al., 1980; Menzel, 1974). Stable carbon isotope measurements of the
standing stock of organic matter also support a largely marine source for DOM (-20 to
-22 %o)(Druffel et al., 1992; Williams and Gordon, 1970). However, there are little data
[DOC]
M
60
40
20
m
80
0
200
-
400
Depth (m)
600 -
800
1000 0
1200
-
-
~-
Figure 1.1 Depth profiles of dissolved organic carbon (DOC) in the Sargasso Sea (filled diamonds)
and the North Pacific Ocean (open squares) (data from Druffel et al. 1992).
directly linking the DOM present in seawater to a plankton source. For example, DOM
concentrations, in general, are not well correlated with either phytoplankton biomass or
primary productivity (Carlson et al., 1994; Chen et al., 1995; Morris and Foster, 1971),
indicating that other processes such as bacterial degradation of DOM must be important
in maintaining steady state DOM concentrations.
Other possible sources of DOM are riverine DOM, which is expected to be
intrinsically unreactive, and thus contribute disproportionately to the marine pool (Ertel et
al., 1986), and input of DOM from sediment porewaters (Chen and Bada, 1989;
McCorkle et al., 1985). The stable carbon isotope data and analysis of lignin oxidation
products in seawater (Meyers-Schulte and Hedges, 1986) indicate that terrestrial DOM is
not a large faction of the total DOM dissolved in seawater. However, the stable carbon
isotope data alone cannot be used to exclude riverborne DOM as a possible source, due to
the presence of terrestrially derived DOM with heavy stable carbon isotopic values (e.g.
C4 plants have 613C values between -10 and -18 %0) (Fry and Sherr, 1984).
1.3
Cycle of DOM in the Ocean
The cycle of DOM in the ocean has been studied using radiocarbon analyses. The
A 4 CI values of DOM isolated from the deep ocean range between -400 9%(4000
radiocarbon years) in the deep Atlantic Ocean to -525 %0(6000 radiocarbon years) in the
Pacific Ocean (as shown in Figure 1.2, (Druffel et al., 1992)). The 2000 year age
difference between the DOM in the deep Atlantic and Pacific Ocean is similar to the
1500 year ocean circulation time derived from measuring the A14C values of DIC. These
14 c12c
1
14
A14C =14C- 2(3C+25 1+
14
C
1000'
;
142
12
14 C
sta dard
14c
(I 12 c)
1
stan dard
X1000
data reveal that (1) a large fraction of DOM escapes degradation on time scales of 6000
years, and (2) DOM must be transported between the two ocean basins by deep-water
circulation. Also shown in Figure 1.2 is the A'4C value of the total DOC in surface
waters of the Atlantic and Pacific Ocean (-150 to -230 %o). These values are more
enriched in
14C
than deep ocean values, but depleted compared to surface ocean DIC
(approximately 130%o; the reported values are for the years 1987-1991). Given the 6000
year age of DOM in the deep Pacific, and the 1500 year ocean circulation time, it is clear
that a large fraction of DOM is cycled over several ocean mixing cycles. The DOC
concentration profile in Figure 1.1 shows that surface waters DOC concentrations are
approximately double the deep-water values. Thus surface waters can be described as an
admixture of refractory, deep ocean DOM (50%) and new, surface derived DOM (50%).
Carrying out a simple mass balance calculation
A4
surfaceDOC
= 0.5 x (- 525)+ 0.5 x (A14 C)neDOM = -200
(the A'4C value of surface waters is expressed as surface DOC, while the unknown A4C
value for the new DOM produced in the surface ocean is expressed as new DOM), shows
that the A 4C value of the new DOM added at the surface is approximately 130 %o,in
good agreement with the A 4 C value of DIC in the surface ocean. This supports the
conclusion that the new component of DOM, added in the surface ocean, is produced
from DIC during photosynthesis.
Depth
(m)
-200% (10
Atl tic
Pacific
Indian
-390% (-94
-
(-221
4500
E
W
Figure 1.2 The A14C values of DOC and DIC (in parentheses) in different parts of the Worlds Ocean.
Arrows depict the direction of ocean circulation.
Using the age of total DOM in the deep ocean and the reservoir size (-600 Gt C),
the flux of DOM out of the ocean can be calculated to be 0.1 Gt C/yr. This flux out can
be supported entirely by the riverine flux into the ocean, and is only a small percentage of
the total annual primary production (50 Gt C/yr). Given the rapid degradation rates of
certain organic compounds and the high heterotrophic bacterioplankton production rates
observed in the ocean (Fuhrman and Azam, 1982), it is clear that certain a fraction of
DOM must cycle on shorter time scales.
(Carlson and Ducklow, 1995) showed that DOM can be sub-divided into at least
three different reservoirs of differing reactivity: a "very reactive" fraction, consisting of
soluble biochemicals, turning over on time scales of hours to days (Jorgensen et al., 1993;
Vaccaro et al., 1968); a "refractory fraction" (6000 years old), which accounts for most of
the measured DOM in the ocean, turning over on several thousand year time scales; and
finally a third fraction, "reactive" DOM, which accounts for the 30-40 4M difference
between surface and deep water DOC concentrations, likely turning over on annual to
decadal timescales. Thus, while the radiocarbon results for the total DOM fraction in the
seawater offers useful insights into the overall cycle of DOM, they clearly miss the more
dynamic cycles that must be exist for different components of the DOM pool.
In order to gain any further insight into the factors that control production and
cycling of DOM we need a better understanding of the major chemical components of
this pool.
1.5
Chemical Structure of DOM
Stuermer and Harvey, (1974) conducted a comprehensive investigation of the
chemical structure of DOM isolated by adsorption on to XAD resins (humic substances).
Shown in Figure 1.3 is a model structure for marine humic substances isolated from
seawater (Gagosian and Stuermer, 1977). These humic substances accounted for
0
_-
0O
r0o-
0
0
NH 2
NH
O
H
OH
O'r
OH
NH
CH20H
CH3
Figure 13 Hypothetical structure for dissolved marine humic substances (Gagosian and Stuermer,
1977)
between 5-15% of the total DOM and were characterized by low aromaticity (compared
to terrestrial humic substances), marine stable carbon isotopic values, carbon to nitrogen
ratios of 8, molecular weights < 700 daltons (73% of the total isolate), and an extremely
complex mixture of amino acids, carbohydrates, lipids (of marine origin) and other
biochemicals. This chemical data led these and other investigators to conclude the much
of the DOM in seawater resulted from the random condensation, oxidation and
polymerization of biochemicals present in seawater as proposed by (Kalle, 1966). Since
XAD isolates account for between 5-15% of the total DOM, there is no a priorireason
why the remaining fraction of DOM accumulating in the ocean should be formed by the
geopolymerization of simple biochemicals.
1.5
Organization of this Thesis
The isolation of DOM by adsorption onto XAD resins chemically fractionates
DOM, and isolates only between 5-15% of the total (Hedges, 1992). The first goal of this
thesis was to isolate and chemically characterize a more representative and larger fraction
of DOM from the marine environment, and to test the theory of geopolymerization by
structurally characterizing DOM from different sites within the ocean. This study used
tangential flow ultrafiltration, which separates molecules based on their size (Benner et
al., 1992; Carlson and Mayer, 1985), to isolate the > 1000 MW (nominal molecular
weight) fraction of DOM (high molecular weight (HMW) DOM) for chemical
characterization by a variety of different analytical methods. The results of this
investigation are presented in Chapter 2, and show that up to 60% of the HWM DOM in
surface waters, and 10-20 % of the HMW DOM in deep waters, consists of structurally
related acylated polysaccharides (APS). These APS are characterized by a nearly fixed
ratio of carbohydrate:acetate:lipid carbon and a very characteristic distribution of seven
monosaccharides.
The IUPAC numbering of monosaccharide carbon atoms, and the two possible
orientations of the anomeric carbon proton (C-1) are shown in Figure 1.3. This anomeric
carbon is particularly important because it is a key linkage site for most carbohydrates,
including APS isolated in this study. Figure 1.4 shows the structures for the 10 different
monosaccharides discussed in Chapter 2.
6
4
2
OH
H
1
3
[3-anomeric
H
a-anomeric
OH
Figure 1.4 Monosaccharide structure: carbon and protons are labeled 1-6 as shown (a), 3 anomer (b)
and a anomer (c).
The discovery that the major biochemicals in HMW DOM were present in a near
constant ratio at all the diverse sites examined in this study is surprising and suggests a
direct biological source (biosynthesis). However, the constant ratio of biochemicals in
HMW DOM does not exclude the goepolymerization model. If DOM consists of
refractory geopolymers with residence times longer than the mixing time of the ocean,
then these geopolymers would exist well mixed throughout the oceanic reservoir. To test
the conclusion of Chapter 2, that a large fraction of "reactive" DOM is from biosynthesis,
we chemically characterized the HWM DOM exudates of three common phytoplankton.
It was found that organic matter with chemical characteristics similar to APS were
present in the exudates of two of the three species of phytoplankton, indicating that the
source of this APS is direct biosynthesis.
OH
CH 2OH
HO
OH
O
CH20H
OH
o
,
lOH
HO
OH
mannose
HO
OH
OH
rhannose
fucose
HO
OH
0
OH
HO
OH
OHOH
OH
ribose
OH
arabinose
xylose
ribose
CH 2OH
HO
OH
OH
OH
---
OH
OH
galactose
glucose
HO
CH 2 0H
OH
mOHn
HO
H
glu
HO
o
NH2
CH 2OH
OO
OH
OH
NH
CH3
N-acetyl
glucosamine
glucosamine
N-acetyl glucosamine
Figure 1.5 Eight neutral monosaccharides identified in HMW DOM: hexoses (A); deoxy hexoses
(B); pentoses (C). Glucosamine in its acetylated and non-acetylated form (D) is also shown.
Finally, we attempted to use the chemical data and radiocarbon measurements to
determine the mechanisms that control the cycling of APS in the ocean. The results from
this investigation, presented in Chapter 4 shows that the individual monosaccharides that
constitute APS are of contemporary origin both in the surface and deep ocean, with Al4 C
values similar to surface DIC. The presence of DOM with contemporary with A'4 C
values in the deep sea indicates recent, autochthonous production. The radiocarbon data
further confirm that the constant biochemical ratio observed for seawater APS do not
result from a well mixed reservoir, but from a ubiquitous marine source as shown in
Chapter 3.
These data require a shift from the currently accepted paradigm that
geopolymerization is the mechanism which forms the accumulating DOM in the ocean.
While it is likely and possible that some fraction of DOM is formed by these
mechanisms, our data suggests that at least 20% of the DOM accumulating in seawater is
produced by direct biosynthesis. Further, contrary to the picture of DOM cycling
generated by using A14C values of total DOM, the data presented in this thesis indicates
that there are more dynamic cycles for certain fractions of DOM within the larger cycle
of passive transport by ocean circulation.
1.6
Goals of this Thesis
The goal of this thesis was to chemically characterize dissolved organic matter
(DOM) isolated by ultrafiltration (20-30% of the total DOM), and use the chemical
information to better understand the cycle of DOM in the ocean. The thesis is divided
into three parts: (1) the chemical characterization of high molecular weight (> 1 k Da)
(HMW) DOM isolated from different oceanic regimes; (2) the use of the chemical data to
identify possible sources of HMW DOM to the marine environment; and (3) the use of
radiocarbon measurements to understand the cycle of HWM DOM in the ocean.
1.7
References
Azam F., Fenchel T., Field J. G., Gray J. S., Meyer-Reil L. A., and Thingstad F. (1983)
The ecological role of water-column microbes in the sea. Marine Ecology
ProgressSeries 10, 257-263.
Bacastow R. and Maier-Reimer E. (1991) Dissolved organic carbon in modeling oceanic
new production. Global Biogeochemical Cycles 5(1), 71-85.
Bada J. L. and Lee C. (1977) Decomposition and alteration of organic compounds
dissolved in seawater. Marine Chemistry 5, 523-534.
Benner R., Pakulski J. D., McCarthy M., Hedges J. I., and Hatcher P. G. (1992) Bulk
chemical characteristics of dissolved organic matter in the ocean. Science 255,
1561-1564.
Blough N. V., Zafiriou O. C., and Bonilla J. (1993) Optical absorption spectra of waters
from the Orinoco River outflow: terrestrial input of colored dissolved organic
matter to the Caribbean. Journalof GeophysicalResearch 98(C2), 2271-2278.
Bruland K. W. (1989) Oceanic zinc speciation: Complexation of zinc by natural organic
ligands in the North Pacific. Limnology and Oceanography34, 267--283.
Carlson C. A. and Ducklow H. W. (1995) Dissolved organic carbon in the upper ocean of
the central equatorial Pacific Ocean, 1992: Daily and finescale vertical variations.
Deep-Sea Research 42(2-3), 639-650.
Carlson C. A., Ducklow H. W., and Michaels A. F. (1994) Annual flux of dissolved
organic carbon from the euphotic zone in the northern Sargasso Sea. Nature 371,
405-408.
Carlson D. J. and Mayer L. M. (1985) Molecular weight distribution of dissolved organic
materials in seawater as determined by ultrafiltration: A re-examination. Marine
Chemistry 16, 155-171.
Chen R. F. and Bada J. L. (1989) Seawater and porewater fluorescence in the Santa
Barbara Basin. GeophysicalResearch Letters 16, 687-690.
Chen R. F., Fry B., Hopkinson C. S., Repeta D. J., and Peltzer E. T. (1995) Dissolved
organic carbon on Georges Bank. ContinentalShelfResearch 16, 409-420.
Coale K. H. and Bruland K. W. (1988) Copper complexation in the North East Pacific.
Limnology and Oceanography33, 1084-1101.
Druffel E. R. M., Williams P. M., Bauer J. E., and Ertel J. R. (1992) Cycling of dissolved
and particulate organic matter in the open ocean. Journalof Geophysical
Research 97(10), 15639-15659.
Duursma D. K. (1963) The production of dissolved organic matter in the sea, as related to
the primary gross production of organic matter. NetherlandsJournalof Sea
Research 2, 85-94.
Ertel J. R., Hedges J. I., Devol A. H., and Richey J. E. (1986) Dissolved humic
substances of the Amazon River system. Limnology and Oceanography31, 739754.
Frew N. M., Goldman J., Dennet M. R., and Johnson A. S. (1990) Impact of
phytoplankton generated surfactants on Air-Sea gas exchange. Journalof
GeophysicalResearch 95(c3), 3337-3352.
Fry B. and Sherr E. B. (1984) 6 13 C measurements as indicators of carbon flow in marine
ecosystems. Contributionsin Marine Science 27, 13-47.
Fuhrman J. A. and Azam F. (1982) Thymidine incorporation as a measure of
heterotrophic bacterial production in marine surface waters: Evaluation and field
results. Marine Biology 66(2), 109-120.
Gagosian R. B. and Stuermer D. H. (1977) The cycling of biogenic compounds and their
diagenetically transformed products in seawater. Marine Chemistry 5, 605-632.
Hagstrom A., Larsson A., Horstedt P., and Normark S. (1979) Frequency of dividing
cells, a new approach to the determination of bacterial growth rates in aquatic
environments. Applied EnvironmentalMicrobiology 37, 805-812.
Hedges J. I. (1992) Global Biogeochemical cycles: progress and problems. Marine
Chemistry 39, 67-93.
Johnson K. H., Gordon R. M., and Coale K. H. (1997) What controls dissolved iron
concentrations in the World Ocean? Marine Chemistry 57, 137-161.
Jorgensen N. O. G., Kroer K., Coffin R. B., Yang X. -H., and Lee C. (1993) Dissolved
free amino acids and DNA as sources of carbon and nitrogen to marine bacteria.
Marine Ecology ProgressSeries 98, 135-148.
Kalle K. (1966) The problem of Gelbstoff in the sea. Oceanographyand MarineBiology,
Annual Review 4, 91-104.
Kirchman D. L., Suzuki Y., Garside C., and Ducklow H. W. (1991) High turnover rates
of dissolved organic carbon during a spring phytoplankton bloom. Nature 352,
612-614.
LaRoche J., Nuzzi R., Waters R., Wyman K., Falkowski P. G., and Wallace D. W. R.
(1997) Brown tide blooms in Long Island's coastal waters linked to interannual
variability in groundwater flow. Global Change Biology 3(3), 397-410.
Mague T. H., Friberg E., Hughes D. J., and Morris I. (1980) Extracellular release of
carbon by marine phytoplankton a physiological approach. Limnology and
Oceanography25(2), 262-279.
Martinez J. and Azam F. (1993) Aminopeptidase acitivity in marine chroococcoid
cyanobacteria. Applied EnvironmentalMicrobiology 59(11), 3701-3707.
McCorkle D. C., Emerson S. R., and Quay P. D. (1985) Stable carbon isotopes in marine
porewaters. Earth andPlanetaryScience Letters 74, 13-26.
Menzel D. W. (1974) Primary productivity, dissolved and particulate organic matter, and
the sites of oxidation of organic matter. In: E. D. Goldberg (Ed.), The Sea (John
Wiley & Sons Ltd.), 659-678.
Meyers-Schulte K. and Hedges J. I. (1986) Molecular evidence for a terrestrial
component of organic matter dissolved in ocean water. Nature 321, 61-63.
Moffet J. W. and Zika R. G. (1987) Solvent extraction of copper acetyllacetonate in
studies of copper (II) speciation in seawater. Marine Chemistry 21, 301-313.
Morris A. W. and Foster P. (1971) The seasonal variation of dissolved organic carbon in
the inshore waters of the Menai Strait in relation to primary production.
Limnology and Oceanography16, 987-989.
Palenik B. and Morel F. M. M. (1990) Amino acid utilization by marine phytoplankton:
A novel mechanism. Limnology and Oceanography35(2), 260-269.
Peltzer E. T. and Hayward N. A. (1996) Spatial and temporal variability of total organic
carbon along 140 W in the Equatorial Pacific Ocean in 1992. Deep Sea Research
Special EqPac43(4-6), 1155-1180.
Rue E. L. and Bruland K. W. (1995) Complexation of iron (III) by natural ligands in the
Central North Pacific as determined by a new competitive ligand
equilibration/adsorptive cathodic stripping voltammetric method. Marine
Chemistry 50(1-2), 117-138.
Smith D. C., Simon M., Alldredge A. L., and Azam F. (1992) Intense hydrolytic enzyme
activities on marine aggregates and implications for rapid particle dissolution.
Nature 359, 139-142.
Stuermer D. H. and Harvey G. R. (1974) Humic substances from seawater. Nature 250,
480-481.
Vaccaro R. F., Hicks S. E., Jannasch H. W., and Carey F. G. (1968) The occurrence and
role of glucose in seawater. Limnology and Oceanography13, 356-360.
Voelker B. M., Morel F. M. M., and Sulzberger B. (1997) Iron redox cycling in surface
waters: effects ofhumic substances and light. EnvironmentalScience and
Technology 31(4), 1004-1011.
Williams P. and Gordon L. (1970) Carbon-13: Carbon-12 ratios in dissolved and
particulate organic matter in the sea. Deep-Sea Research 17(19-27).
2.
Structural Characterization of Marine Dissolved High Molecular Weight
Organic Matter
2.1
Introduction
In order to fully understand the relationship between the production, accumulation and
removal of dissolved organic matter (DOM) in the ocean, it is important to first
investigate its chemical composition. Until the mid 1990's direct chemical analyses of
DOM in seawater typically accounted for less than 15% of total mixture (Hedges, 1992)),
and it was believed that the characterization of the remainder of DOM had been frustrated
by its heteropolycondensate nature (Degens, 1970; Gagosian and Lee, 1981). The study
of DOM has been further impeded by the fact that many techniques that can provide
comprehensive structural information, such as nuclear magnetic resonance spectroscopy
(NMR), mass spectrometry (MS) and infrared spectroscopy (IR), require the technically
difficult isolation of DOM from the much more abundant salts in seawater. As discussed
in the Chapter 1, past studies on DOM have focused on the compounds isolated by the
adsorption of acidified DOM onto non-ionic XAD resins (Gagosian and Stuermer, 1977).
This separation requires large manipulations of the pH, which may alter the DOM and
chemically fractionates the DOM during isolation. The recent use of tangential flow
ultrafiltration has allowed investigators to isolate DOM from seawater based on
molecular size, and requires no chemical manipulation of the sample. Using filters with a
nominal cutoff of 1 nm (> 1000 daltons) several investigators have shown that up to 30%
of the total DOM can be isolated from seawater (Carlson and Mayer 1985; Benner et al.
1992; Santschi et al., 1995) using this method. Bulk chemical characteristics of this high
molecular weight (HMW) DOM show that contrary to what was expected (from the XAD
resin isolates) a large fraction of HMW DOM consists of recognizable biochemicals.
Using ' 3CNMR for example, Benner et al. (1992) showed carbohydrate resonances were
the major fraction of HMW organic carbon in surface waters (54%) and significant
fraction in deep waters (25%). In fact, several investigators in the past, using a variety of
different methods, also showed that up to 20% of the total carbon dissolved in seawater
can be identified as polysaccharides (Burney et al., 1979; Ittekot et al., 1981; Mopper,
1977; Sakugawa and Handa, 1983; Sakugawa and Handa, 1985b). In accordance with
these findings, a large body of work has emerged over the last several years showing that
17-50% of the carbon in HMW DOM, and thus at least 7-15% of the total DOC, is
carbohydrate (Aluwihare et al., 1997; Borch and Kirchman, 1997; McCarthy et al., 1996;
Skoog and Benner, 1997).
The large concentration of polysaccharides in HMW DOM would suggest that it
is nitrogen poor, however C/N ratios in both the surface and deep waters are
approximately 15 (Benner et al., 1992; McCarthy et al., 1996), suggesting that HMW
DOM must contain nitrogen. Using '5 NNMR, McCarthy et al., (1998) showed that the
major nitrogen resonance in HMW DOM was amide, arising from proteins or acetylated
amino sugars. In accordance with the carbohydrate data, identifiable biochemicals, for
example bacterial membrane proteins (Tanoue et al., 1995) have also been isolated from
the HMW dissolved organic nitrogen pool. However, amino acid analysis accounts for
only 4-5% of the total carbon in HMW DOM and approximately 11-29% of the total
nitrogen in HMW DOM (McCarthy et al. 1996). Acetylated amino sugars have also been
identified in HMW DOM by GC (Aluwihare et al., 1997), HPLC (Skoog and Benner,
1997) and mass spectrometry (Klap, 1997). Quantification by the two former methods
show that acetylated amino sugars do not account for more than 10% of the total carbon
in HMW DOM.
The purpose of this chapter was to identify the major carbon and nitrogen containing components of DOM isolated by ultrafiltration and to investigate if the HMW
DOM isolated from different parts of the ocean was chemically distinct. The chapter
begins with a synopsis of the sampling sites, and sampling methods, and includes data
from 'H, '3 C and '5 N NMR, monosaccharide composition and some carbohydrate
linkage, lipid, and amino acid analyses.
2.2
Materials and Methods
2.2.1
Sampling Site
In order to investigate the hypothesis that DOM from different locations is
compositionally similar, samples spanning diverse oceanographic environments were
collected and analyzed (Figure 2.1). A suite of samples from the Mid Atlantic Bight were
collected during the Ocean Margins Program sponsored by the Department of Energy. In
addition, two samples from the Northeast Pacific Ocean (Station M), and one sample
each from the Eastern North Atlantic, Hawaii and Woods Hole (Coastal Research
Laboratory) were also analyzed. Sample locations, depths and sampling time are listed in
Table 2.1 (see results section 2.3.1).
2.2.2
Sample Collection and Isolation
All surface samples were collected using a Teflon lined hose, and pumped
onboard ship with a pneumatic, teflon-lined diaphragm pump (Lutz Pumps, Inc.,
Norcross, GA). Deep samples were collected via Niskin bottles. Seawater samples were
45.
- -"Woods
40
Hole
* MAB941
*eMB944
. MABB943
MAS1
.
* MAB947
MAB966,MAB967
35,
30.
60 -
---- 75-------70
65
---------------r
- --
------,
I..
-"
. ,,"N
' -
-
'
40
P961, P96 1600m)
MAB
20
Hawaii
0 ................ 1
-160
-180
-..............
-140
-120
100
-80
-60
40
-20
0
20
Figure 2.1 Map showing the locations of the different samples analyzed for this thesis
pre-filtered through both pre-combusted (4500 C, 12 hours) 293mm, Whatman GF/F (0.7
pm) or Gelman A/E (1.0 pmu)filters and acid (0.1 N HC1) rinsed, 0.2 pm, Gelman
Criticap 100 capsules. Filters were changed when the flow rates decreased. Pre filtered
samples were stored (< 5 h) in Teflon coated barrels (Fluoropure, Chaska, MN) until they
were ultrafiltered. All tubing and fittings were Teflon except for the 273 mm filter holder
(polycarbonate) and the Criticap capsule filters (polycarbonate).
2.2.2.1 Ultrafiltration
Ultrafiltration (the generic term for cross flow filtration and tangential flow
filtration (Carlson and Mayer, 1985), is the only practical technique for processing the
large volume samples (10-1000 1)required for the isolation of DOM for more
sophisticated analytical techniques such as NMR, MS and infrared spectroscopy (IR). In
ultrafiltration, a pre-filtered sample solution flows parallel to the ultrafiltration
membrane, and hydrostatic pressure drives solutes with an effective molecular size less
than the membrane pore diameter, through the membrane (permeate) (Buesseler et al.,
1996). The remaining solution (retentate), containing some of the smaller solutes and all
of the bigger solutes, is transported along the membrane surface and recycled through the
retentate reservoir. Thus the molecules which are rejected by the membrane (i.e. bigger
than the membrane pore diameter) are increasingly concentrated in the retentate over time
(Buesseler et al. 1996).
A schematic for the system used in this study is shown below (Figure 2.2). An
Amicon DC- 10OL ultrafiltration unit was used in this study, and was routinely run with
two spiral-wound polysulphone membrane cartridges, stacked in series. These
membranes have a 1 nm pore diameter and are designed to have a nominal molecular
weight cutoff of 1000 daltons (1 k Da) (Millipore, Bedford, MA; Amicon S1ONI). Thus
these membranes retain compounds whose molecular weights are greater than IkD. With
Surd
Seawater
Valve
urr
"Sterilc"Scar aer in
I
wFluid
0-30 psi
Gauge
Cocentrated fluid
Systn
Two Spial
Reserv r
wound. >IkD
cartrdges
Permeate o t
c60 psi
Teflon Lined
Drr
Valve
Fluoropure Barrel
(closed)
Pump
Amicon DC-IOL Ultrnnfiltrtion Unit
(Equipped with two Amicon Spiral
Wound Polysulphone Membrane
Cartridges)
Figure 2.2 Schematic representation of the procedure for the isolation of HMW DOM
two cartridges in series, the Amicon system can be run with a 20 psi (pounds per square
inch) pressure differential between the inlet and the outlet. At no time during
ultrafiltration was the inlet pressure allowed to exceed 50 psi, and thus the outlet pressure
never exceeded 30 psi. Samples were delivered from the teflon-lined barrels to the
ultrafiltration system by applying slight air pressure to the barrels. In this manner 10s to
100s of liters of seawater were processed continuously. A 200 1surface sample of
seawater typically took 8-12 h to concentrate to approximately 2 1. Unless samples were
desalted by diafiltration (described in section 2.2.2.2) on site (Hawaii and CRL),
concentrated samples were frozen (-4" C) in pre-combusted glass bottles, or acid rinsed
(concentrated HC1) teflon bottles. Directly after processing each sample, the
ultrafiltration unit was then rinsed with 4 1of Milli Q water, which was collected and
frozen as above.
Membranes were calibrated using two compounds with molecular weights close
to the membrane pore diameter, maltoheptaose (Sigma, St. Louis, MO), molecular weight
1.15 k Da, and Gramicidin S (Sigma, St. Louis, MO), molecular weight 1.2 k Da.
Samples were dissolved in Milli Q water (DOC concentrations 10-19 gM), at a
concentration of 1 mg C/ 1. The concentration factor in all cases was 4. Recoveries were
determined by measuring the DOC concentration in sample prior to ultrafiltration and in
the retentate following ultrafiltration. Purity of the samples following ultrafiltration was
also determined using 'HNMR spectroscopy. Retention characteristics were 80-±10% and
50 % for the carbohydrate and protein, respectively. Other investigators have found
higher values, typically 80% for vitamin B-12 (1.2 k Da) and > 93 % for molecules larger
than 6 k Da (Guo and Santschi, 1996; Gustafsson et al., 1996).
2.2.2.1.1
Cleaning The Membranes
In all cases, water other than seawater, was collected from a Millipore Milli Q
system, (four cartridge system withl pre-filter cartridge, a carbon removal cartridge
(Super C), and two ion exchange cartridges (Ion X)). Water was collected after the
resistivity increased beyond 18 ohms and was filtered through a 0.2 pm filter prior to
collection. The water isolated directly from the system had a DOC blank of 15 iM. New
cartridges, and cartridges that were stored for over three weeks were thoroughly cleaned
before use. Cartridges were cleaned by sequentially passing 4 1 liters of Micro detergent
(1%), 0.01 N sodium hydroxide (NaOH) and hydrochloric acid (HC1) solutions in Milli Q
water through the system (Guo et al., 1994). Between each cleaning solution, the
cartridges were rinsed with 20 1 of Milli Q water. The pH of the permeate was measured
after the last Milli Q rinse to insure that it was between 6-7 units (that of unprocessed
Milli Q water). Between samples, on board ship, cartridges were only cleaned with base
and acid solutions (MAB941 through MAB947, MAB961), except when a deep sample
followed a surface sample (before MAB966, MAB967, and between P961 and P962)
when all three cleaning solutions were used. Before ultrafiltering each sample, 3-4 1 of
seawater were used to condition the cartridges.
2.2.2.2 Diafiltration
Diafiltration or desalting was performed in the laboratory (except in the case of
Hawaii and CRL) on thawed samples. During diafiltration, the sample was added to the
ultrafiltration reservoir along with 2 1 of Milli Q water and ultrafiltered to between 1.5-2
1. Following this, another 2 1of Milli Q water was added and re-concentrated to 1.5 1.
This procedure was repeated until the permeate from the last 2 1rinse gave no visible
precipitate when added to a saturated solution of silver nitrate (0.5g/50 ml). In most
cases 10 dilutions (for a total of 20 1) were used in diafiltering each sample. Following
diafiltration, samples were concentrated to approximately 2 1and stored at -200 C in
glass or teflon bottles for a few weeks to several months.
Prior to analysis, all samples were lyophilized and stored at room temperature or 40 C.
2.2.2.3 Desalting
Prior to the linkage analysis, samples (CRL, Hawaii and ENL) were further
desalted using a cation exchange resin. Lyophilized samples were applied, dissolved in 1
ml of Milli Q water (5 mg sample/2 ml of resin bed), to the cation exchange resin (AG
50W-X8, BioRad Laboratories, Hercules, CA) in the hydrogen form, and eluted with
water (3 column bed volumes). The resin was rinsed with 3x3 ml of Milli Q water prior
to packing. Columns were packed in 6" combusted glass pipettes (6 or 8 mm I. D.),
using Milli Q water.
Mass balances after desalting showed that material was lost to the column, but in
all cases removal of salt could account for this loss. Three samples from the deep ocean
(UF967, UF948, and UF9412 (750 m)), in addition to the three samples mentioned
above, were tested to determine changes in the C/N ratios incurred during desalting.
Samples containing between 1-10% carbon (C/N ratios between 1 to 4) showed an order
of magnitude change in their C/N ratios following desalting, while samples with > 25% C
showed only a 20% increase in their C/N ratio. Neither charged sugar (galactouronic
acid and glucosamine) standards nor Bovine Serum Albumin (BSA) showed losses to the
column during desalting. Recovery of monosaccharides and changes in the distribution
of monosaccharides following desalting are discussed more fully in Chapter 4.
2.2.3
Dissolved Organic Carbon (DOC) Analysis
Total DOC concentrations were measured by filtering 20-40 ml of seawater
through pre-combusted GF/F filters (47 mm, Whatman) and collecting the filtrate in precombusted glass vials fitted with teflon-lined caps. Any surface that came into contact
with seawater was teflon, stainless steal or glass. High molecular weight (HMW) DOC
concentrations were obtained by directly sampling the retentate and diluting with a
known volume of Milli Q water. All samples were acidified with 50% phosphoric acid
H3PO 4 (0.5 M) immediately after collection and stored at room temperature or 4°C.
Samples were analyzed for DOC by high temperature catalytic oxidation (Peltzer and
Brewer, 1993) after purging with nitrogen to remove inorganic carbon. Precision for
triplicate analyses were better than +3.5% and the upper limit for the blank was 25 tM
based on the DOC concentration of acidified distilled water.
2.2.4
Elemental Analysis
DOC samples were analyzed for carbon, hydrogen and nitrogen content using an
EA 1108 elemental analyzer with Eager 200 data acquisition software (Fisons
Instruments, Inc., Beverly, MA). Between 300-1000 pg of lyophilized DOM was
weighed into 8 x 6 mm tin cups using a Sartorius Micro balance, and folded with forceps
into very small packages prior to analysis. Samples were run with blank cups and known
standards in order to correct for the C and N associated with the tin cups.
2.2.5
Nuclear Magnetic Resonance Spectroscopy
2.2.5.1 General Principles
Nuclear magnetic resonance spectroscopy (NMR) is a non-destructive technique
for the determination of chemical structure and conformation. Today, with the
development of magnets with high field strengths (> 750 MHz) and multi-dimensional
techniques, NMR is used to routinely analyze the primary, secondary and tertiary
structures of complex chemicals.
NMR is based on the principle that the magnetic moment of nuclei with spin
numbers of V2 (for example 'H, ' 3 C and '5 N) will align with or against an externally
applied magnetic field. At equilibrium, according to the Boltzmann distribution, there
are more nuclei aligned with the applied magnetic field, than against. The nuclei aligned
with the magnetic field are in the lower energy state (spin +2) and those aligned against
the magnetic field are in the higher energy state (spin -V2). In a NMR experiment, the
nuclei of interest are excited to a higher energy state by the application of a radio
frequency (rf) pulse at right angles to the applied magnetic field. This deviation from the
Boltzmann distribution, resulting from the rf pulse, changes the direction of the net
magnetization vector of the nuclei from its orientation with the applied magnetic field
(the positive z axis by convention) to an orientation in the x-y plane. Once the rf pulse
turns off, the individual nuclei relax back to their ground state at a rate that is dependent
on the immediate chemical environment of each nucleus. This relaxation, or loss of
energy to the environment, is enabled by two processes known as longitudinal relaxation
(whereby a nucleus dissipates its energy to surrounding molecules) and transverse
relaxation (whereby energy is dissipated among nuclei). As the nuclei relax back to their
ground state (alignment with the z-axis) the magnitude of the magnetization vector in the
x-y plane begins to decrease. This decrease is detected by the receiver (which measures
the size of the vector in the x-y plane) and is referred to as the free induction decay (FID).
The FID, which is a function of time, is Fourier transformed to yield peaks as a function
of frequency. Thus the x-axis of the NMR spectrum shows peaks as a function of
frequency.
The difference in the absorption position of a particular proton from the
absorption position of a reference proton is called its chemical shift. One of the major
and more easily understood factors that influence chemical shift is electronegativity. For
example, since carbon is more electropositive than hydrogen, the sequence of proton
adsorbtions in the series CH 4, RCH3, R2CH 2 and R3CH is from lower frequency to higher
frequency (and higher field to lower field). Furthermore, the chemical shift of protons in
RCOCH 3 is downfield (lower field) from RxCHx, while the non-alcohol proton of
RCHOH, absorbs even further downfield (higher chemical shift). These latter chemical
shifts are caused by the influence of the oxygen atom, which draws the electron density
away from the proton nucleus, thus decreasing its shielding by electrons.
Units of 8 (in p.p.m) are usually employed to report chemical shifts because they
are dimensionless and are independent of the applied magnetic field. Higher 8 values
correspond to higher frequencies and lower field.
2.2.5.2 Experimental Conditions for Proton NMR ('HNMR)
All spectra were acquired using a Bruker AC300, 300 MHz spectrometer (Bruker
Instruments, Inc., Manning Park, Billerica, MA) in conjunction with an Aspect 3000 data
system (Spectrospin AG, Industriestrasse 26, CH-8117 Faellanden, Switzerland). The
quantity of sample used for each acquisition depended on the solubility of the material
and the spectrometer availability. Typically, 10 mg of DOM was used with acquisition
times ranging between 1-2 days. Spectra can be acquired on much smaller quantities of
material (< 100 p.g), but much longer acquisition times are necessary for these
experiments. All samples were dissolved in 750 pl tol ml of >99.9 atom % purity D20
(Aldrich Chemical Company, Milwaukee, WI) and acquired at 2980 K. In most cases
samples were first dissolved in 1-2 ml of 99.9 atom % D20 and freeze-dried. This was
repeated 1-2 times in order to insure that the exchangeable protons in the sample were
saturated with deuterium. In all cases, chemical shifts are reported in the standard "6"
notation (units of p.p.m), relative to HDO (6 = 4.8 p.p.m). Acquisition parameters for a
typical IHNMR spectrum are given in the Appendix. The resonance from HDO
(deuterated water) was suppressed in all cases (irradiation power was between 1.5
x 10-
and 1.5 x10 4 Watts). All samples were run in 5 mm emperor glass NMR tubes (Wilmad
Glass Co., Buena, N.J.), and transferred using combusted glass NMR pipettes (Wilmad
Glass Co.). NMR tubes were reused and cleaned between each sample by soaking in a
2:1 solution of concentrated HNO 3 and HC1, followed by rinsing with copious quantities
of tap water, distilled water and Milli Q water, and oven drying at 60 OC.
The area under each peak in the NMR spectrum (in solution state) is proportional
to the number of nuclei in the sample which absorb at that frequency. Thus, it is possible
to determine the relative ratios of the different proton functional groups that are present in
the sample by comparing the areas under each peak. In order to obtain the relative areas
spectra were photocopied and each peak was weighed out in triplicate.
2.2.5.3 ' 3 C and iSNNMR
Both carbon and nitrogen NMR are based on the same general principles
discussed in 2.2.5.1. However, greater amounts of sample and/or much longer
acquisition times are necessary to obtain a
abundance of the
13C
13C
NMR spectrum, since the natural
atom is 1.1% of all carbon atoms, and the sensitivity of the 13C
nucleus for NMR is 1.6% that of the 1H nucleus. Due to the large amount of sample
needed,
13CNMR
and 15NNMR were performed on samples isolated in Woods Hole only
(CRL).
For
13CNMR,
up to 100-200 mg of samples were prepared as above and finally
dissolved in D2 0 (chemical shifts are reported relative to carboxylic acid (176 p.p.m)).
Due to the large quantity of sample needed, all spectra were obtained in 10 mm tubes
(Wilmad Glass Company, Buena, NJ). The
13CNMR
spectrum shown in this chapter was
acquired using the Bruker AC300, 300 MHz spectrometer described above, in 27 000
scans, using a pulse width of 4 ms, and a recycle delay of 2 ms, with the decoupler on
(0.15 Watts) for 100 gs.
In the case of 15N NMR, the sample was analyzed in solid state. The natural
abundance of 'N is 0.37% of all nitrogen atoms, and its relative NMR sensitivity for
equal numbers of nuclei is only 1/1000 times that of protons, making natural abundance
' 5N NMR often impractical because of the long acquisition times necessary to obtain
useful spectra (Gust et al., 1975). However, the greater frequency separation of nitrogen
resonances as compared to both hydrogen and carbon (Gust et al., 1975) makes Nitrogen
NMR a very useful tool for determining the forms of nitrogen present in a sample.
Solid state spectra were acquired by Dr. Anthony Bielecki of Bruker (Billerica,
MA), using 240 - 400 mg of sample (3% nitrogen). Solid state NMR is typically
performed on approximately 300 mg of material. Spectra were acquired with a Bruker
Avance 400 MHz NMR instrument in 250, 000 scans, using a spin rate of 5000 MHz. In
this chapter chemical shifts for 15NNMR are reported relative to liquid ammonia.
In this study, 15NNMR spectra were acquired for two samples of HMW DOM. In
one case, 420 mg of HMW DOM (13 mg of nitrogen) isolated from Woods Hole
seawater (November, 1998) as described in section 2.2.2.1, using a combination of both
polysulphone and regenerated cellulose membrane cartridge filters (both from Millipore),
were submitted for 'SNNMR analysis. In the second case 1.3 g sample, containing 39 mg
of nitrogen, (ultrafiltered as above) was hydrolyzed in 1 M HCl at 105 0 C for 17 h and
lyophilized prior to acquiring the 15NNMR. In the latter case, 15 mg of pyridine was
added as an external standard (206 p.p.m) prior to data acquisition.
2.2.6
Carbohydrate Analysis
2.2.6.1 Monosaccharide Analysis
The monosaccharide composition of HMW DOM was determined according to
York et al. (1985). Hydrolysis of the HMW DOM was carried out by heating a mixture of
300 - 700 gg of dry sample in 0.5 ml of 2 M trifluoroacetic acid (TFA), containing 20 jtg
of myo-inositol as an internal standard, at 1210 C for 2 h in a tightly capped glass vial.
Samples were cooled to room temperature and the acid was removed by drying under a
stream of nitrogen. Monosaccharides were reduced to alditol acetates using 250 iil of a
solution of sodium borohydride (NaBH4 ) dissolved (10 mg/ml) in 1 M NH 40H, and
incubated at room temperature for a minimum of 2-3 h. The reaction was quenched with
glacial acetic acid (250 .l) to decompose the excess borohydride. Glacial acetic acid and
trace amounts of water were removed by adding methanol (3x250 pl) and then drying
under a stream of nitrogen. The dry alditols were per-acetylated with 100 pl of acetic
anhydride and 20 p1 of 1-methyl imidazole and mixed thoroughly. After 15 minutes at
room temperature, the excess acetic anhydride was quenched with 0.5 ml of Milli Q
water. After a further 10 minutes at room temperature, 0.5 ml of dichloromethane was
added. The solution was vortex mixed, and the alditol acetates were extracted into the
lower, dichloromethane layer. This layer was withdrawn and dried over Na2 SO 4
(anhydrous), transferred and the dichloromethane was removed under a stream of
nitrogen. The alditol acetates (Figure 2.3) were then re-dissolved in 100 pl of methanol
and analyzed using an HP 5890 gas chromatograph equipped with a 30 m, DB-5 fused
silica column (0.25 mm ID, and 0.20 pm film), using a temperature program of
150(1)/2/240(15), (initial temperature (time)/ramp/final temperature (time)).
CH 2OAc
OH
O
NaBH
OH
HO
OH
4
Ac20O
OAc
3
AcO
OAc
OH
CH2OAc
Figure 23 Synthesis of alditol acetates from monosaccharides
2.2.6.2 Linkage Analysis
Glycosyl linkage analysis was performed on the HMW DOM by preparing
partially methylated alditol acetates according to the method of Hakomori (1964), with a
few modifications (Figure 2.4). Freeze dried samples were first desalted as described in
section 2.1.3. Up to 1 mg of the desalted, dry sample was dissolved in 0.5 ml of dry
dimethylsulphoxide (DMSO), under nitrogen with stirring. Methylation of the sample
was initiated by adding 0.5 ml of dry 2M potassium dimethyl sulphinyl (dimsyl) anion
prepared earlier by adding 4.8 g of potassium hydride (the Hakomori procedure uses
NaH) to 10 ml of dry DMSO.
Following the addition of the dimsyl anion, the reaction tube was flushed with
nitrogen, capped, and stirred. After 8 h the sample tube was cooled in an ice bath until
frozen, and 1.0 ml of iodomethane was slowly added to complete the methylation. The
tube was then vortex-mixed gently, capped, and stirred overnight at room temperature.
Following this, 0.5 ml of Milli Q water was added, and the tube vortex-mixed. Finally,
the reaction mixture was evaporated under a stream of nitrogen at room temperature until
all the methyl iodide was removed (i.e. until the solution was clear and homogeneous).
The methylated HMW DOM was purified by solid phase extraction using a C-18 reverse
phase cartridge (Waters Sep-Pak). The cartridge was preconditioned with 8 ml of
acetonitrile, followed by 8 ml of Milli Q water. The methylated sample was applied
slowly to the top of the cartridge, washed with 8 ml of Milli Q water, and eluted with 2-3
ml of acetonitrile. (For Hawaii and CRL, 1HNMR spectra of methylated polysaccharides
were obtained prior to hydrolysis).
The dry, permethylated polysaccharide was hydrolyzed at 121 C for 2 hr in 0.5
ml of 2M trifluoroacetic acid containing 20 tg of myo-inositol. The reduction and
acetylation of the generated monosaccharides were carried out as described above for
alditol acetates, but substituting sodium borodeuteridefor sodium borohydridein the
reduction step. The partially methylated alditol acetates, finally dissolved in 100 pl of
methanol, were analyzed by GC-MS, using a VG-Autospec Mass Spectrometer equipped
with a HP 5890 Gas Chromatograph. Samples were run on a Supelco, SP2330 30 m
OMe
OH
DMSO/KH
HO
1
10
HO
o
HO
HO
RO
MeO Me
OH
O
HO
OH
0
MeO"
HO
Mel
.
3
O
-OR
L1
OMe
MeO
O, Me
M eO,
2oM
ROW
-OR
.
3
2
RO 2
0
OMe
2 OH
TFA
NaBD 4
OMe
MeOO
MeO MeO
1
MeO
OMe
HO_6OH
+ MeO
HO
HO
CHOH
IMeO
2
D
3
'. CHOH
I
D
~
OMe
OH
CHOH
HO
I
D
Pyridine
AC20
CHDOAC
CHDOAc
HDOAc
-OMe
+
-OMe
OAc-
1
Terminal(1)
Me
OAc
OMe
-OAc
CH,OMe
-OAc
+
-OAc
F
CH,OMe
2
Branched (1,3,4)
OAc
CH2 OMe
3
Non Branched
Figure 2.4 The reaction scheme for the preparation of partially methylated alditol acetates for
linkage analysis of polysaccharides.
fused silica column (0.25 mm ID, and 0.20 pm film), using a temperature program of
80(2)/30/170/4/240(10) (temperature (hold time)/ramp rate/temperature/ramp
rate/temperature (hold time)) with Helium as the carrier gas. Methylated
monosaccharides were identified by comparing chromatographic and mass spectral
properties with known standards.
2.2.7
Lipid Analysis
High molecular weight DOM (8-10 mg C in 10 -15 ml) samples were analyzed
for the presence of bound fatty acids by acid hydrolysis (HCl (4 N, 4-22 h at 100*C)) and
saponification (KOH (20%, 24 h at 100 0 C)) (Sidorczyk et al., 1983). HMW DOM
samples (8-10 mg C in 10 -15 ml) were also analyzed for the presence of ether bound
lipids with boron tribromide (BBr 3, Aldrich) (90 0 C for 4h) according to Benton and
Dillon (1942). All reactions were performed in sealed tubes under nitrogen. All
glassware was combusted and extracted with methanol and dichloromethane. Acids,
bases, and samples dissolved in water were pre extracted with ethyl ether or
dichloromethane to remove any free lipids. The organic extracts were analyzed by
'HNMR to insure that samples and reagents were free of lipids. Following hydrolysis,
the pH of the aqueous layer was adjusted to 3-4 units (using NaOH), and extracted with
chloroform or diethyl ether (for the ester bound lipids) and hexane (for the ether bound
lipids). The extracts were dried (by rotary evaporation, at room temperature) and redissolved in deuterated methylene chloride for 'HNMR. Prior to gas chromatography,
dry extracts were derivatized using N,O-bis [trimethylsilyl] trifluoroacetamide (BSTFA,
Pierce, Rockford, IL) in pyridine (heat at 600 C for 15 min), for fatty acid analysis and
derivatized with acetic anhydride and pyridine (room temperature, 15 min) for alcohol
analysis. Samples were then dried and re-dissolved in dichloromethane and separated on
a DB-5 (J&W Scientific, Folsom, CA) column using a temperature program of
55(5)/4/320(10).
After drying the extracts, samples were re-dissolved in water and analyzed for low
molecular weight fatty acids (lipids) using ion chromatography (samples were analyzed
by Dr. J. Seewald, WHOI, using a Dionex Ion Chromatograph with suppressed
conductivity detection).
2.2.8
Amino Acid Analysis.
Amino acids were analyzed by high pressure liquid chromatography (HPLC) using
o-phthaldialdehyde derivatives, by Dr. S. Pantoja (according to Pantoja et al., 1997).
Four hydrolysis methods were compared in this study. Two mg of CRL HMW
DOM (containing 60 jtg of nitrogen), were hydrolyzed using: (a) 2M TFA, for 2 h at
1210 C, (b) 1N HC1 for 11 h at 105 0 C, (c) 7N HC1 for 1.5 h at 150 0 C and (d) 4N HC1 for
17 h at 100 0 C. Following hydrolysis, samples were neutralized with sodium hydroxide,
prior to O-phthaldialdehyde derivatization. Amino acids were separated by HPLC using
a Beckman Ultrasphere C18 column (5 pm, 250x4.6 pm), and detected by fluorescence
using a Hitachi 1000 Fluorometer (with excitation and emission wavelengths set at 330
nm and 450 nm respectively). Individual amino acids were separated and eluted in 35
minutes using a gradient of sodium acetate (25 mM, pH 5.7. with 2% tetrahydrofuran)
and methanol (18-100%), with a flow rate of 1 ml/min.
2.3
Results
2.3.1
Ancillary Data
The ancillary data for each of the samples discussed in this chapter are shown in
Table 2.1. As shown in this table, the samples analyzed in this study spanned a range of
DOC concentrations (40 -210 ipM). Carbon to nitrogen ratios in the HMW DOM
samples ranged from 11 to 21. This range is larger than that previously observed by
other investigators (15-22 (Benner et al., 1992; McCarthy et al., 1996)), but the samples
in Table 2.1, span more diverse environments. The average C/N value for this data set
(not including UF945) was 15.
Two numbers for the recovery of HMW DOC from total DOC are given for each
sample. The first number, (% HMWl), is the recovery ofHMW DOC (from total DOC)
after ultrafiltration, calculated using the concentration of HMW DOM in the retentate).
The second number is the recovery ofHMW DOC after ultrafiltration, diafiltration and
lyophilization (%HMW2). Typically, ultrafiltration isolates between 20-30% of the total
DOC (Benner et al., 1992; Carlson and Mayer, 1985). The average amount of HMW
DOC isolated in this study was approximately 26% in accordance with earlier studies (%
HMWl). However, marked losses in HMW DOC occurred following diafiltration, with
>50% loss of carbon for most samples. Except in the case of MAB961, the samples
which were most efficiently diafiltered (i.e. samples that contained >20% carbon) showed
the highest losses of carbon. Other investigators have also noted carbon losses during
diafiltration, for example, Guo and Santschi (1996) observed that 35-39% of the HMW
carbon was lost during diafiltration. These investigators suggest that the
Sample
Depth
Latitude Longitude Date
MAB941
MAB943
MAB944
MAB945
MAB947
MAB948
MAB961
MAB966
MAB967
P961
P962
ENA
Hawaii
Surface
Surface
Surface
Surface
Surface
750 m
Surface
300 m
750 m
Surface
1600 m
Surface
600 m
40.33
38.7
39.3
38.93
37.18
37.18
36.78
36.7
36.7
34.83
34.83
51.6
19.67
72.03
73.15
75.5
75.06
74.27
74.27
75.87
74.58
74.58
123.15
123.15
40.11
156
Apr-94
Apr-94
Apr-94
Apr-94
Apr-94
Apr-94
Jul-96
Jul-96
Jul-96
Oct-96
Oct-96
Apr/Jun-96
Jan-95
DOC(piM)
CIN
%C
89
72
209
132
121
47
160
40
40
72
40
200
80
14
16.5
20.4
2
21
13.8
12
11.4
6.2
7.3
12.3
8.5
3.29
37
28
28.7
26.3
17.3
16.1
12.4
12.8
12
11
10.3
HMW DOM
(mg)
51.4
41.0
39.5
29.1
222.0
20.0
26.9
38.4
28.9
N/D
N/D
%HMWI
%HMW2
22
37
21
37
26
13
18
10.7
17.1
33
22
27
16.5
7.6
23.1
4.2
5.6
9.1
5.6
Table 2.1 Ancillary data for each of the samples analyzed as part of this thesis.
The prefixes for each of the samples are as follows: MAB is the Mid-Atlantic Bight; P is the North East Pacific Ocean; ENA is the Eastern North Atlantic (a
sample from the Sint Annaland salt marsh in the Oosterschelde estuary (Dr. V. Klap), and Hawaii was filtered at the Natural Energy Laboratory in Hawaii.
%HMW 1 was calculated from the concentration of the retentate following ultrafiltration. % HMW2 was calculated using the HMW DOM (mg) yield after
ultrafiltration, diafiltration with 20 1of Milli Q water, and lyophilization. %C and C/N are reported for HMW DOM (after ultrafiltration and diafiltration)
loss of material during diafiltration may be caused by the incomplete ultrafiltration of
LMW DOM.
Riverine and coastal samples were highly colored (clear, bright yellow), while
deep samples showed more color than open ocean surface samples (suggesting photo
bleaching of surface HMW DOM). Lyophilized samples were powdery and colors
ranged between white and yellow.
2.3.2
Nuclear Magnetic Resonance Data
2.3.2.1 Proton NMR
The major resonances present in HMW DOM spectra were from carbohydrates (4
- 5.5 p.p.m. (anomeric), 3 - 4.0 p.p.m (CHOH), and 1.3 p.p.m (CH)). Resonances for
acetate (2.0 p.p.m (CH 3COO/N)) and low-molecular weight, non-acetate lipids were also
present (0.9 p.p.m (CH 3) and 1.3 p.p.m (CH 2)) (Figure 2.5). The peak centered at 4.8
p.p.m in all the 'HNMR spectra arises from partially deuterated water (HDO). Proton
NMR spectroscopy of HMW DOM showed similar spectra for all surface ocean samples
(Figure 2.6 and Figure 2.7A).
Given the monosaccharide data that are discussed in section 2.3.3, the lipid data
discussed in section 2.3.4, and the area under each of the major peaks identified above,
the relative proportion of major biochemicals in surface water HMW DOM can be
calculated. The monosaccharide data are used determine the contribution of deoxy
sugars (rhamnose and fucose) to the area of the lipid peak at 1.3 p.p.m. The
monosaccharide data presented in section 2.3.3 show that deoxy sugars are 30% of the
total neutral monosaccharides in each of the samples. Thus 30% of the peak area
between 3 - 4 p.p.m comes from the deoxy sugars. Since 3 out of the 8 non-alcohol
protons in deoxy sugars give rise to lipid resonances, the contribution of deoxy sugars to
the total lipid peak (at 1.3 p.p.m) can be further calculated.
Since the carbon to hydrogen ratios of these biochemicals are known (C/H ratios
are 1 carbohydrates, 0.67 for acetate and 0.5 for lipids), the measured areas in the
1HNMR
spectra may be expressed in terms of carbon. Table 2.2 shows results from the
integration for 20 surface seawater samples (including all the samples listed in Table 2.1).
H2 OH
HO
H
H
H3
O R
HOH
1 0H
H
HH
H
8.0
H
7.0
6.0
4.0
5.0
3.0
2.0
1.0
PPM
Figure 2.5 The major 1HNMR resonances identified in HMW DOM. The baseline for each of the
major peaks and the whole spectrum are shown on the figure as dashed lines.
As can be seen from Figure 2.6, the resonances from the major biochemicals identified
above were superimposed on a featureless background of overlapping resonances. The
amount of HMW carbon contained within the background was calculated using the C/H
ratios listed above, and measuring the baseline area under each of the three major peaks.
According to the data in Table 2.2, identifiable biochemicals, carbohydrates, acetate and
lipid, accounted for 70% of the total carbon in the sample. Thus the data shown in Table
2.2 can be recast to determine the contribution of each of the three major biochemicals to
the total identifiable carbon. This ratio is 80-4:10±2:9±4 for carbohydrate:acetate:lipid
carbon and stays remarkably constant in the surface water HMW DOM samples analyzed
thus far (see Figure 2.6 and 2.7 A).
Carbohydrate C
Surface Water(n=20) 56±4
Deep Water(n=8)
25±4
Acetate C Lipid C
6±4
7±2
2±0.5
7±3
Baseline C
30±3
64t7
62
4
2
32
Deep Water*(n=2)
* Refers to ratios after the sample was desalted using a cation exchange resin.
Table 2.2 Relative abundance (in carbon) of the major biochemicals identified in HMW DOM by
'HNMR
Shown in Figure 2.7 B is a typical 'HNMR spectrum of HMW DOM isolated
from the deep ocean (P962). Comparing the spectra shown in Figure 2.7 A and 2.7 B it is
apparent that all the major resonances identified in surface seawater (spectrum A) are still
present, including acetate. However, the relative proportions of these biochemicals are
quite different from surface water, as shown by the data in Table 2.2. The major
differences between the two spectra are a substantial decrease in the acetate and
carbohydrate peaks relative to the lipid peak and a relative increase in the area under the
base line.
A. North Atlantic
HDO
\A
B. North Pacific
I
I
8.0
7.0
I
6.0
I
I
I
5.0
. . .. I . .. . . .. .
I
4.0
3.0
2.0
.
1.0
PPM
Figure 2.6 Two 'HNMR spectra for surface HMW DOM isolated from the North Atlantic (MAB943)
(A) and North Pacific (B) (P961). Note the similarities in the resonances present and the relative
proportion of each resonance.
The carbohydrate: lipid ratio decreased from 9 in surface waters to 4 in deep water while
the ratio of carbohydrate:acetate increased from 8 to 12.5.
As shown in Table 2.2, the
relative contribution of uncharacterized carbon, calculated using the background area in
1HNMR
spectra from the deep ocean, doubled compared to surface values (64% (deep)
vs. 30% (surface). This data indicates that the component containing carbohydrate and
acetate decreases with depth, or that the background component increases with depth.
This depth related decrease in the relative fraction of carbohydrates is consistent with the
'3 CNMR data of Benner et al. (1992).
A. Surface Ocean
I
B. Deep Ocean
C. Deep Ocean#
. . . . . . . . . . . . . .
6.0
5.0
3.0
4.0
. . . . . . . . . . . . . .
2:0
. . . . . . , . . . .
1.0
PPM
Figure 2.7 The figure shows 1HNMR spectra for a typical surface ocean sample (A) (CRL), a typical
deep ocean sample (B) (P962) and a spectrum seen for two of the eight deep samples analyzed (C)
(MAB944). Dashed lines in Spectrum B show the baseline for each of the identified peaks.
Much of the relative increase in background resonances (seen in Figure 2.7C)
resides between 1.0 and 2. 5 p.p.m, indicating a relative increase in methyl and methylene
protons.
Thus, HMW DOM isolated from the deep ocean appears to contain a larger
relative amount of reduced compounds. For both surface and deep HMW DOM samples
the baseline contains a "smear" of resonances (i.e. a variety of resonances that are not
well resolved). As discussed in section 2.2.5.1, as the excited proton nuclei relax back to
the ground state they dissipate energy at a rate that is dependent on the immediate
chemical environment of each nucleus. The "smear" of resonances in the baseline
indicate that the sample contains several protons with a slightly different chemical
environment. Molecules that are coiled or molecules whose functional groups are closely
associated may give rise to a variety of poorly resolved resonances. A mixture of closely
related molecules or large molecules that rotate very slowly in solution may also give rise
to poorly resolved resonances.
Neither surface nor deep ocean spectra show significant contribution from
aromatic protons (< 1% of the carbon in the samples).
The third spectrum shown in Figure 2.7 C is one that has been observed for two of
the eight deep samples examined by 'HNMR spectroscopy (a sample from a depth of 750
m in the MAB (35.5 0 N, 74.70 W), and a sample from 1000 m in the Sargasso Sea
(36.5 0N, 65.5 0 W)). Although carbohydrate, acetate and lipid resonances are present in
this spectrum, additional resonances (compared to Figure 2.7 A and B) are apparent in the
region 1.9 - 2.8 p.p.m. When Figure 2.7 C is compared to the spectrum shown in Figure
2.9, which shows a mixture of two standards, maltoheptose and Gramicidin S (protein),
some of the same resonances are apparent. For example, the resonances in the region 1 3.0 p.p.m in the standard mixture arise entirely from Gramicidin S. Spectra similar to
Figure 2.7C have also been identified in culture work (see Chapter 3).
Figure 2.8 shows the spectrum of a deep ocean sample (MAB948) before
desalting (A), after desalting (B) and following hydrolysis with 4M HCI (as in section
2.2.7). The spectrum of the desalted sample resembles the surface seawater samples
6.0
6.0
50
5.0
4!0
PPM
30
4.0
3 0
20
2 0
10
1 0
PPM
Figure 2.8 1HNMR spectra of MAB948, a HMW DOM sample from a depth of 750 m before
desalting (A), after desalting with a cation exchange resin (B) and after hydrolysis (C)
more closely than Figure 2.8A, and clearly shows that both the surface and deep samples
contain the same resonances, but that deep samples have a much more pronounced
baseline. In spectrum 2.8 B, the acetate peak is more easily identified and the
carbohydrate peak is relatively more pronounced, however, no overall loss (< 4%) of
carbon is incurred between 2.8A and 2.8B. The spectrum shown in Figure 2.8 C
illustrates that hydrolysis of both the acetate and carbohydrate fractions yield similar
results to those observed for HMW DOM isolated from the surface ocean.
Maltoheptaose and Gramicidin (5:1)
6.0
5.0
4.0
PPM
3.0
2.0
1.0
Figure 2.9 Proton NMR spectrum of a 5:1 mixture of maltoheptose and Gramicidin S, in D 20.
Resonances enclosed in brackets arise from proteins.
2.3.2.2 Carbon NMR
The individual carbon resonances in HMW DOM can also be observed directly
using ' 3 C NMR. A 13C NMR spectrum of surface seawater HMW DOM (CRL) is shown
in Figure 2.12. The major resonances arise from carbohydrates (20 - 50 p.p.m (CH 3), 60
- 85 p.p.m (CHOH), 85 - 115 p.p.m (anomeric)), acetate (20-50 p.p.m (CH 3), 160-180
p.p.m (COO)) and lipids (20 - 50 p.p.m (CH 3 , CH 2)). As expected, the anomeric carbons
from carbohydrates fall in the region centered at100 p.p.m, and the secondary
carbohydrate peak at 65 p.p.m likely arises from C6 of non-deoxy hexoses. Aromatic and
unsaturated carbons give rise to resonances between 120-140 p.p.m. In our ' 3CNMR
spectra there are no clear resonances in this region.
The relative proportions of the major biochemicals calculated from Figure 2.10
are 64:11:25 (carbohydrate:carboxylic:lipid). If the area under the lipid and carbohydrate
peaks are adjusted according to the contribution from deoxy sugars (deoxy sugars are
30% of the carbon in the sample, and the C6 resonance (in the lipid region) is 1/6 of this),
the relative ratios change to 67:11:22 (carbohydrate:carboxylic:lipid). The ratio of
carbohydrate to acetate carbon determined from both 'HNMR and the lipid analysis is
8L2. Using this ratio, the amount of acetate carbon resonances in the sample can be
calculated (8). Half of this acetate carbon is buried in the lipid resonances, and thus the
carbon distribution in HMW DOM, determined by ' 3 CNMR spectroscopy, can be further
expressed as 65:8:18 (carbohydrate:acetate:lipid). The ratio of identifiable
carbon:baseline carbon in this spectrum is 60:40. Benner et al., (1992) found similar
carbohydrate:carboxylic:lipid carbon ratios using ' 3 CNMR spectroscopy for HMW DOM
isolated from the surface Pacific Ocean (54:13:26).
C2,C3,C4,C
5
r--
CH 2 , CH 3
0
R
O-
180
Figure 2.10
13
160
140
0
120
100
PPM
CNMR spectrum of HMW DOM in D2 0
80
I
60
I
40
2
20
Both the 'HNMR and the 13CNMR data (presented in this thesis and present in the
literature) are in good agreement and show that the major biochemicals identified by
NMR are present in a relatively fixed ratio throughout the ocean. It is the initial data
obtained by NMR that led the way for the other analyses presented in this chapter, and
highlights the utility of NMR spectroscopy in characterizing even very complex mixtures
of biochemicals.
2.3.2.3 Nitrogen NMR
In order to determine the major types of nitrogen present, HMW DOM was also
analyzed by '5 N NMR spectroscopy. Shown in Figure 2.11 A is the solid state 15NNMR
spectrum of CRL HMW DOM. The spectrum was dominated by one peak at 122 p.p.m
(92% of the total area) belonging to amide nitrogen (CNCO). Two other peaks were also
present in this spectrum: a spinning sideband at -23 p.p.m from the peak at 122 p.p.m,
and a peak at 40 p.p.m from free amines (CNH 2, or C2NH, 8% of the total area).
The amide resonance likely arises from the peptide bond of proteins or from Nacetylated sugars (such as N-acetylated glucosamine or galactosamine). The amine arises
from free amines in proteins (terminal amino acids, and amino acids such as arganine and
lysine which have two or more nitrogen atoms) and/or amino sugars (glucosamine, and
galactosamine). In order to differentiate between the amide resonances from Nacetylated sugars and other amide resonances (such as peptide bonds), the DOM sample
was hydrolyzed with IM HC1 for 17 hours at 105 0 C. Proton NMR confirmed the
hydrolysis of the acetate group (Figure 2.12 A and B), indicating that any N-acetylated
sugars were converted to amino sugars.
0
-
R
NH2
RNH 2, R2NH
Spinning Sideband
300
350
250
200
150
100
50
0
B
-50
RNH2,R2NH
/
Pyridine
I
I
350
300
R
I
250
I
I
200
150
NH2
I
100
50
I
0
I
-50
Figure 2.11 Solid state ISNNMR spectra of HMW DOM, before (A) and after (B) 1 M HCI
hydrolysis for 11 h at 105*C. Resonances are reported relative to NH 3. Pyridine was added as an
external standard to quantify the total nitrogen detected by 1SN NMR
A. Pre Hydrolysis
HDO
No Acetate Peak
B. Post Hydrolysis and Freeze Drying
HDO
.
8.0
7.0
6.0
5.0
I
4.0
3.0
2.0
"
VII
I
1.0
PPM
0
Figure 2.12 1H NMR spectra of HMW DOM, before (A) 1 N HCI hydrolysis for 11 h at 105 C and
after hydrolysis and freeze drying to remove acetic acid (B).
The hydrolyzed sample should contain no amide resonances if all the amide
signals in the ' 5N NMR spectrum arise from N-acetylated sugars. A ' 5N NMR spectrum
of the sample after hydrolysis is shown in Figure 2.12 B. The resonance at 204 p.p.m
arises from 0.95 mmoles of pyridine added as an internal standard to calculate the
recovery of nitrogen (103%). In contrast to the unhydrolyzed sample, the major
resonance in the hydrolyzed sample is at 40 p.p.m, previously identified as the amine
region. This peak accounts for 67% of the total nitrogen in the sample (not including
pyridine). Some amide resonances remain after hydrolysis (120 p.p.m) and they
represent the remaining 33% of the total nitrogen resonances in the sample (not including
pyridine). The 'HNMR (Figure 2.12) confirms that all the acetate is hydrolyzed, insuring
that the remaining amide signal does not arise from N-acetylated sugars. Thus the
remaining amide nitrogen likely arises from unhydrolyzed proteins in the sample.
However, we cannot rule out the possibility that the remaining amide in sample may arise
from a yet unidentified type of amide nitrogen.
In order to determine the origin of the amine nitrogen, the hydrolyzed sample was
analyzed for the presence of amino acids (by Dr. Silvio Pantoja, WHOI) and N-acetylated
sugars. Amino acid analysis was able to account for 15-18% of the nitrogen in the
hydrolyzed sample. This yield of amino acids is in accordance with literature values (1128% (McCarthy et al., 1996)). OPA derivatization (Section 2.2.8) of the sample also
showed the presence of glucosamine which accounted for 5% of the total nitrogen.
Further hydrolysis of the CRL sample (Section 2.2.8) by various methods (including 4M
HC1, for 19 hours at 100 0 C) did not increase the yield of either amino acids or amino
sugars. Thus the remaining 73 - 80% of the amine signal in the NMR spectrum was not
directly accounted for as amino acids or amino sugars.
2.3.3
Lipid Results
Hydrolysis of HMW DOM caused a change in the 1HNMR of each sample
(Figure 2.8C, Figure 2.13 B), especially in the region of 2.0 p.p.m. Following the
hydrolysis, fatty acids were extracted from the aqueous phase (section 2.2.7) and
derivatized for GC analysis. No fatty acids were detected by GC. Ether cleavage with
BBr 3 yielded no lipids either; however this analysis was only performed on one sample
and may give different results for other samples. Proton NMR of the organic layer after
extraction (Figure 2.13 D) showed one resonance at 2.0 p.p.m (not including peaks from
the solvent (diethyl ether)). The extracted aqueous layer (Figure 2.13 C) showed the
removal of the peak at 2.0 p.p.m. The lack of other peaks at 0.9 and 1.3 p.p.m (CH 3 and
CH2 from lipids) in the 1HNMR spectrum of the organic extract, coupled with the
chemical shift of the single peak, indicated the presence of acetic acid, which must be
present within HMW DOM as acetate. The absence of CH 3 and CH 2 resonances in the
organic layer confirms the absence of fatty acids detected by GC, since the temperature
program was designed to detect fatty acids of chain lengths greater than sixteen carbons.
The presence of acetic acid in the organic layer was further confirmed by ion
chromatography. The amount of acetic acid determined by ion chromatography (9±2.7
pmoles) was not statistically distinguishable from the amount of acetic acid expected
from the 1HNMR spectrum (10.8 tmoles).
R
R-N
A. Pre Hydrolysis
CH3
O
B. Post Hydrolysis
HO
CH 3
C. Post Extraction
D. Organic Layer After Extraction
Diethyl Ether
6.0
1.0
PPM
Figure 2.13 1H NMR of HMW DOM prior to 4 N HCI hydrolysis (A), following hydrolysis (B), the
aqueous layer (pH = 3) after extraction with ethyl ether (C) and the ethyl ether layer (D).
2.3.4
Monosaccharide Analyses
The HMW DOM samples were hydrolyzed and analyzed for monosaccharides as
discussed in Section 2.2.6.1 (Figure 2.14) and the relative mole percent of
monosaccharides in each of the samples is shown in Table 2.3. All samples showed the
presence of the same neutral monosaccharides. Galactose was observed to be the most
abundant monosaccharide, with glucose, mannose, xylose, rhamnose and fucose being
only slightly less abundant, and arabinose usually being the least abundant neutral
monosaccharide. Analysis of the some of the samples (MAB941, MAB943, MAB944,
MAB945, and MAB947) for uronic acids and amino sugars showed that these sugars
comprised <12 % of the total monosaccharides. In the case of surface samples
(MAB961, MAB965, CRL, P961, ENL and Hawaii), the neutral monosaccharides
comprised between 15-20 % of the total carbon in the sample. Deep ocean samples
(MAB964, MAB967 and P962) contained smaller amounts of neutral monosaccharides
(between 4-9% of the total HMW DOM carbon). This depth associated decrease in the
total monosaccharides coincides with the 'HNMR data which shows decreasing
carbohydrate resonances with depth. The monosaccharides yields observed in this study
are comparable with existing data (6-15% in surface waters vs. 1.6-3.5% in deep water
(Borch and Kirchman, 1997; McCarthy et al., 1996; Skoog and Benner, 1997).
The data in Table 2.3 are plotted in Figure 2.15 to better illustrate the mole
percent monosaccharide distribution. Despite the differences in the total amount of
monosaccharide between surface and deep waters, the relative distribution of
Arabinose
PA:
200
175
Surface Pacific (P961)
150
125
1oo.
75
25t
A
A
I
Figure 2.14 Raw gas chromatograms of alditol acetates.
Sugars
Rhamnose
Fucose
Arabinose
Xylose
Mannose
Galactose
Glucose
MAB941
0.16
0.15
0.05
0.13
0.14
0.19
0.19
MAB943
0.16
0.17
0.05
0.15
0.12
0.21
0.14
MAB944
0.18
0.16
0.08
0.15
0.12
0.15
0.16
MAB945
0.17
0.14
0.09
0.11
0.18
0.22
0.09
Sugars
Rhamnose
Fucose
Arabinose
Xylose
Mannose
Galactose
Glucose
MAB964
0.17
0.20
0.09
0.13
0.12
0.18
0.12
MAB965
0.14
0.16
0.09
0.11
0.15
0.23
0.12
MAB967
0.18
0.16
0.09
0.10
0.17
0.16
0.14
CRL96
0.12
0.16
0.05
0.15
0.15
0.22
0.15
0.16
0.14
0.15
0.25
MAB947
0.17
0.16
0.08
0.15
0.13
0.18
0.13
MAB961
0.20
0.17
0.07
0.11
0.13
0.21
0.10
MAB961
0.19
0.17
0.08
0.11
0.13
0.22
0.10
ENA
0.14
0.17
0.08
0.15
0.15
0.18
0.13
Hawaii
0.10
0.15
0.09
0.13
0.17
0.21
0.16
P961
0.13
0.17
0.09
0.11
0.14
0.23
0.13
P962
0.17
0.18
0.07
0.08
0.15
0.20
0.15
MAB946
0.18
0.13
Table 2. 3 The relative distribution of neutral monosaccharides (as analyzed by alditol acetates) in HMW DOM.
All samples (except MAB964 and MAB965) are listed in table 2.1. MAB965 is the surface counterpart of MAB966 and MAB967. MAB964 was taken from a
0
depth of 750 M at 35.5 0N and 74.7 W. MAB94 1, MAB943, MAB944, MAB945 and MAB947 were also analyzed for uronic acid and amino sugars. In. all
cases, these sugars were 10% of the total sugars in each sample, except in the case of mab945 where they accounted for 12% of the sample. Glucose distributions
varied ±25% of their mean distribution, arabinose distributions varied ±20%, and all other sugars varied between ±13-15%. Replicate monosaccharide analyses
varied ±5%.
P962
P961
..
.
..
-...
..
..
..
...
..
Hawaii
ENA
CRL96
...........
.......
.......
MAB967
".
!i:!:i:i!:::::::::::::
i~i ! !iiii~iiiiiiii
ili~ii~iii!
-- m
MAB965
MAB964
MAB961
':,iii~iiiiiiiii~iii~.
:::
---------
.
,..
...
. ............
...
.....i~
.....................
.
..
...
...
..........
........
. . ...............
:::::j:
MAB947
MAB946
MAB945
MAB944
MAB943
MAB941
0.2
0.4
0.6
0.8
Relative Mole Percent
o Rhamnose
EFucose
8Arabinose
EXylose
* Mannose
N Galactose
8 Glucose
Figure 2.15 The mole percent distribution of the seven neutral monosaccharides identified in HMW
DOM.
monosaccharides is remarkably similar for all samples and compares well to distributions
observed by other investigators (McCarthy et al., 1996; Sakugawa and Handa, 1985b).
In order to determine how closely the monosaccharide distributions in each
sample were related a correlation coefficient analysis was performed on the samples
shown in Figure 2.16. The results of the analysis are given in Table 2.4, and as an
example, the results for the extent of correlation of all the samples with MAB941 are
shown plotted in Figure 2.16. Most samples correlate well with MAB941, showing the
close relationship between the monosaccharide distributions in HMW DOM isolated
from diverse locations. The sugar distribution of P962 and MAB967 are closely
correlated (>0.9). These are both deep ocean samples, but their sugar distributions are
poorly correlated with MAB964, which is also a deep ocean station. However, MAB964
had high monosaccharide yields (9% of HMW DOM) and its iHNMR was unique
(Figure 2.7C). In fact MAB964 is not well correlated with any of the samples (except
MAB944). The sugar distribution in the Hawaii sample is closely correlated (0.85) with
P961, which is also a Pacific Ocean surface sample. The correlation coefficient data for
the MAB samples shown in Table 2.4 correspond well to the direction of water flow
established for this region (Figure 2.17) (Beardsley et al. 1976). The correlation
coefficient data for this region demonstrates that although the monosaccharide
distributions among samples are closely correlated, there are subtle differences that
differentiate samples in this region driven by the advection of DOM by different water
masses. This is consistent with the DOM concentrations which increase southward along
the shelf of the MAB indicating that at each location the DOM consists of DOM advected
along the shelf, DOM produced at that location and/or DOM brought in by rivers (e.g.
MAB 95, MAB961 and MAB967).
If the correlation coefficient analysis is repeated without glucose the degree of
correlation increases between most samples. The greater variability in glucose compared
to other monosaccharides is not surprising given the importance of glucose in energy
storage polysaccharides. Thus while the other neutral monosaccharides likely occur
together in a single polysaccharide, a larger fraction of glucose is present in other
polysaccharides as well.
MAB941
MAB941
MAB943
MAB944
MAB945
MAB947
ENA
MAB961
P961
P962
MAB965
MAB964
CRL
HAWAII
MAB967
MAB941
MAB943
MAB944
MAB945
MAB947
ENA
MAB961
P961
P962
MAB965
MAB964
CRL
HAWAII
MAB967
1
0.84
0.79
0.47
0.78
0.85
0.59
0.7
0.78
0.65
0.56
0.85
0.7
0.67
MAB943
0.84
1
0.77
0.64
0.95
0.96
0.83
0.85
0.76
0.82
0.44
0.94
0.68
0.63
MA944
MAB945 MAB947 MAB961
MAB964
0.47
0.79
0.78
0.59
0.56
0.77
0.64
0.95
0.83
0.44
1
0.26
0.87
0.62
0.8
0.26
1
0.66
0.84
-0.09
0.87
0.66
1
0.9
0.52
0.62
0.65
0.84
0.72
0.31
0.62
0.84
0.9
1
0.33
0.41
0.75
0.72
0.79
0.23
0.63
0.69
0.76
0.83
0.57
0.37
0.87
0.73
0.86
0.14
0.8
-0.09
0.62
0.33
1
0.62
0.62
0.82
0.66
0.32
0.22
0.57
0.46
0.42
0.08
0.6
0.76
0.73
0.81
0.51
MAB965 MAB967 CRL
0.65
0.67
0.81
0.63
0.37
0.6
0.87
0.76
0.73
0.73
0.85
0.57
0.86
0.81
0.98
0.65
0.84
0.92
1
0.72
0.14
0.51
0.8
0.56
0.78
0.44
0.72
1
ENA
0.86
0.94
0.62
0.62
0.82
0.99
0.66
0.85
0.68
0.8
0.32
1
0.87
0.56
0.85
0.96
0.62
0.65
0.84
1
0.72
0.9
0.72
0.85
0.31
0.99
0.85
0.57
HAWAII
P961
0.7
0.68
0.22
0.57
0.46
0.85
0.42
0.85
0.6
0.78
0.08
0.87
1
0.44
P962
0.7
0.85
0.41
0.75
0.72
0.9
0.79
1
0.84
0.98
0.23
0.85
0.85
0.65
0.78
0.76
0.63
0.69
0.76
0.72
0.83
0.84
1
0.85
0.57
0.68
0.6
0.92
Table 2.4 Results from the correlation coefficient analysis of the mole percent monosaccharide
distributions in different HMW DOM samples.
0.9
0.8
0.7
-0.6
- 0.5
g 0.4
0.2-
<
LUC
51
0.1
0
2
4
6
8
10
12
14
Figure 2.16 Correlation coefficient data for the mole percent monosaccharide distribution of each
HMW DOM sample. All correlations are reported relative to MAB941
_--l
..r -~---lr------ ---- iI__._
cr_.-.--ll^l.r- I-r~.-ll- ...~-y-
45
-
40
Woods Hole
AB941
MAB94
VAB
S-AB943
MABOM
-
35
AB947
MAB965,MAB966,MAB967
..
........
..............
-.............
.-......--.....
.........-......
........-...
.......
.....
.....
...
..
......
3 0 .....-..........
............
..
....
- -....
-80
-75
-70
-65
Figure 2.17 A schematic representation of the flow of water in the MAB constructed using the
correlation coefficient data for the mole percent monosaccharide distributions.
A standard mixture of seven monosaccharides was made using the
monosaccharide distributions determined by the alditol acetate analyses, and its 'HNMR
is shown in Figure 2.18 A. Also shown in Figure 2.18 is a 'H NMR spectrum of a
hydrolyzed HMW DOM sample (B). The similarities between these two spectra are
remarkable and confirm that much of the DOM in surface water is carbohydrate.
i
However, there is a discrepancy between the monosaccharide yields determined from the
alditol acetate data and the 'HNMR. The combined hydrolysis yields of neutral and
charged monosaccharides do not exceed 30%. However, NMR data estimate that
carbohydrates account for between 50-60% of the total carbon in HWM DOM. Thus,
either biochemicals other than carbohydrates give rise to the resonances in the region 34.5 p.p.m in the 'HNMR spectrum or HMW DOM contains carbohydrates (neutral and
charged) that are not hydrolyzed or identified by conventional analytical methods.
23.5
Methylation Analysis
Several attempts were made to methylate the polysaccharides in HMW DOM
according to the method in section 2.2.6.2 (UF941, UF947, CRL, Hawaii). Following
methylation, polysaccharides were extracted into dichloromethane and analyzed by
'HNMR. Proton NMR spectra of each of the methylated samples were very similar.
Shown in Figure 2.19 are spectra for an aqueous sample prior to methylation (A) and the
organic extract of methylated CRL (B). The R-O-CH 3 groups of the methylated
carbohydrates give resonances at 3.0 p.p.m. There are some large peaks in the
methylated sample near this region (2.8 p.p.m) that may be the peaks of interest.
However, the high ratio of anomeric (4.5 to 6.0 p.p.m) to non-anomeric (3 to 4.5 p.p.m)
carbohydrate protons, and the large peaks between land 2 p.p.m suggest that the sample
contains other compounds besides polysaccharides. If polysaccharides
A. Monosaccharide Mixture
CH3
HDO
0I
__ __
I_
____
B.
Hydrolyzed
Surface
Water DOM
B. Hydrolyzed Surface Water DOM
6.0
4.0
PPM
2.0
1.0
Figure 2.18 The 1H NMR spectrum of a standard mixture of monosaccharides (A) and a hydrolyzed
HMW DOM sample (B).
A. Pre-Methylation
B. Organic Layer After Methylation
0
H
"'OH
Au
6.0
4.0
PPM
3.0
2.0
1.0
Figure 2.19 1EN MR spectra of HMW DOM prior to methylation (A) and following methylation and
extraction into an organic solvent (B).
are present in these extracts, even in small concentrations, the NMR data imply that the
majority of anomeric hydrogen atoms are in the a orientation (5.0-6.0 p.p.m).
This methylation reaction should be performed using CD 3I to determine which of
the peaks in Figure 2.19 B arise from the reaction of polysaccharides and other
compounds with the CD 3 group. It may also be important to methylate the sample
several times to insure that polysaccharides are completely methylated and to desalt the
sample as thoroughly as possible prior to methylation. If performed correctly this is
likely to be the best method for isolating the carbohydrate fraction from the total HMW
DOM.
23.6
Linkage Analysis
A typical total ion trace of partially methylated alditol acetates (PMAA) from
HMW DOM is shown in Figure 2.20. PMAA's were identified by their mass spectra and
retention time. Shown in Figure 2.22 A, B and C are the mass spectra of four different
linkage types identified in each sample. Accompanying each spectrum is the typical
fragmentation pattern expected from a particular compound. Some common fragments of
PMAA's are m/z 118 and 101/102 peaks that can be generated as shown in Figure 2.21.
Other characteristic peaks are m/z 129 for 1,2 -linked sugars, and m/z/ 131 for deoxy
hexoses (Figure 2.21B and C) .
Also given in the appendix of this chapter is a list of the retention times of known
standards that are necessary to make the correct peak identifications.
o
.4
0%
0t
I%0
to
1
0
®
O
C-
W
W
r
40
0
in
n
1
In
e~
8-
58
17:38
17:24
17:01
16:16:0
-----
16:42
16:26
16:14 16:06
/O
^
.44
15:48
cc15
15:33
15:19
15_
Y
14:58
039
14:24
114
1J
8
14-
141
14:01
13:51
,
13:32
13:19
12: 1
E
0J
W
cc
03
'in1
WQ
+
ae
Ik
Ea
11:3
E-m
Figure 2.20: The total ion current of partially methylated alditol acetates from a HMW DOM
sample. Each peak was identified using mass spectral data and retention time data.
A. Terminal Rhamnose
i ^~~__n
O _)_~OL ____~_lr
InnL
~_ 10:11. rr ^_ ~I-rm.l
utoSc
cEO EI0 MlagnetBpI285140 TIC 6583897 FIagS:HALL
Sample Text:CRL1
10
°so
118
.2.9E5
95.
.2.75E
90.
.2.6ES
.2.4E5
80
,13
2.3ES5
1
75.
.2.1ES
70
2.0E5
65
0
,1.9E5
60
.1.7ES
1.65ES
50
1.4ES
Sso
45
.1.3ES
40
1.1E5
35
.1.055
I.0ES
162
175
8.6E4
25
7.1E4
r
20.
,o
15
5.7E4
4.3E4
;o
10
2.9E4
-
!D
140
145
150
155
1 0
Is
1710
CHDOAc
118
Meo162
131
,
MeO
MeO-H3
H3C
OMe
OAc
Ac
175
1.4E4
0
61
5
180
15
10
~cm S0.00
195
m/z
B. Terminal Mannose
ib_ JU-VJUWin IUUUPPMAcq.15-OV-1 6 10:44:15+b:4J
File:
1040 Ident:9
LtoSpeCEQEl.+Hagnet BpM:145Bpl:120768TIC:1542156Flag.:ALL
SampleText:CRL1
CA1:ULU4A
1.2ES
100
102
.
90.
118
1.125
1.2
101-
129
.5
1.05E
85
80.
9.7E4
75-
9.1E4
70.
8.5E4
65s
7.8E4
60.
7.2E4
s
so
161
162
87
6.6E4
6.0E4
205
45
5.4=4
4.8E4
40
4.2E4
35.
3.6E4
o
30
25
3.0E4
20
2.4E4
1.8E4
15
10
1.2E4
-
5t) 0
70
0
90
0
1
1 0 160
1
110
80
190
200
0
220
2 0
2 0
2 0
2 0
102
118
MeO----/
1
205
MOMe
161
OAc
CH2OMe
101
129
87
45
145
0
0
2 O
m/z
C. 1,4 Galactose
File:LNI07IC Ident:b5 654 Her Det 0.2b S:4 Acq:22-JAN-1998 17:41:28 +18:27 Ca:LNII07A
AutoSpecEQ EI+ Magnet BpM:118 BpI:445875 TIC:6875778 Flags:HALL
Sample Text:EH17 Me File Text:1.Oul Inj. from MeOH
100
4.5E5
95.
4.2E5
90o
118
4.0E5
85
.3.8E5
3.6E5
.3.3E5
3.1E5
.2.9E5
60
2.7E5
55-
2.5E5
50
2.2E5
""
129
2.OE5
233
1.8E5
35
1. 6E5
30.
1.3E5
25.
1.1ES
162
20.
.8.9E4
189
15
16.7E4
H 0
10
54.5E4
5
--
W
S 80 10 120 10 10 10 20
22
CD
2.224
!
0
3
030303
46
4
S46" 1"
.o
A o
k n0O
n'
~'
n
a
0
CHDOAc
MeO
118
OMe
189
233
-----
145
AcO-
CH2OMe
CAc
129
Figure 2.21 Mass spectra of terminally linked rhamnose (A), terminally linked mannose (B), 1,4linked galactose (C).
m/z
The linkage patterns for three different HMW DOM samples isolated from the
surface ocean are shown in Figure 2.22. The abundance of each type of linkage is
expressed as a percent of the total linkages present for each sugar. Linkage type "T"
refers to monosaccharides linked only at carbon 1, while linkage type 1,3 refers to
monosaccharides linked at both carbon 1 and carbon 3, etc.
Of the pentoses, only arabinose is present as both the pyranose (p) (six membered
ring) and the furanose (f) (five membered ring) form. Arabinose (f) was only present at a
terminal site. In most cases, the most abundant type of linkage was the terminally linked
monosaccharide. The large percentage of terminal linkages may be an artifact of the
analysis. If the polysaccharide is only partially methylated, then the sites that are most
accessible (i.e. monosaccharides present at the terminus of the polymer) will be
methylated preferentially.
As shown in the Figure 2.22, each monosaccharide is linked at several different
sites. Glucose has the fewest different linkages; for example, the HMW DOM sample
ENA contains only terminally linked glucose. Furthermore, glucose only appears at the
terminus of a polymer or linked at 3 or more positions. All other monosaccharides show
a variety of different linkages.
It is clear from Figure 2.22 that linkage patterns are complex and distinct in the
three samples analyzed in this study. The use of linkage analysis is imperative in
understanding and comparing the structure of polysaccharides isolated from these
I
1,2,3,4
1,2,3,4,6
1,2,4
1,2,4,6
1,3,4
1,2,3,6
I
1,4,6
1,3,6
1,4
r
1,3
T
1,3,4
1,2,3,4
1,4
1,2,4
|
1,6
1,4
1,3
1,3
1,2
T
1,2
i
T
1,2,3,4,6
S1,2,3,4
W
F 1,3,4,6
1,3,4
1,2,4
1,2,3,6
1,2,3
1,2,4,6
1,4
1,3
1,2,3,4,6
-
MMMMOMEMMMMMMME
1,2
1,2,4,6
T
1,2,3,6
1,2,3,4
1,2,4
1,2,4
1,2,3
1,3,4
1,6
1,2,3
1,3
1,3
1,2
1,2
T
T
a
W
VI00h
-4
2A
0
0
0ga
Figure 2.22 The relative distribution of different types of linkages per sugar for three different HMW
DOM samples.
different sites. Thus, it is important to determine if the differences exhibited in the
linkage patterns are real or an artifact of the experimental method. This is the first time
an analysis of this type has been performed on such a complex natural substrate and as
such, the method needs refinement to be more certain of the its results.
2.3.7
Amino Acid Analysis
The relative distribution of amino acids in CRL determined in this study are
shown in Table 2.4 (for CRL) along with the average distribution of amino acids in
seawater HMW DOM (McCarthy et al.1996).
Different acid strengths, temperatures
and time of hydrolysis had no systematic effect on the yield of amino acids but affected
the relative distribution of amino acids in the sample (as illustrated by the standard
deviation data). In all cases, total amino acid yields did not exceed 20% of the total
nitrogen and 4-5% of the total HMW carbon. CRL1, CRL2, CRL3 and CRL4 were all
initially hydrolyzed with 1 M HCI at 105 OC for 11 hours. Following this hydrolysis,
CRL2 was hydrolyzed with 2M TFA at 121 0 C for 2 h (the standard hydrolysis method
used in this thesis), CRL3 was hydrolyzed with 7 N HC1 at 150 OC for 1.5 h (the
hydrolysis method used by McCarthy et al. (1997)), and CRL4 was hydrolyzed with 4 N
HCl at 100 °C for 17 h (the standard method used to liberate muramic acid from
peptidoglycans (Discussion section)). The variation in amino acid distributions resulting
from different hydrolysis methods is unexpected. These differences may reflect
degradation of amino acids during certain hydrolysis procedures. Amino acid standards
were not tested for each method.
acid
Amino acid
Amino
AA
Aspartic Acid
Glutamic Acid
Serine and
Histidine
Glycine
Threonine and
Arginine
Alanine
Tyrosine
Methionine
Valine
Ph-Alanine
Isoleucine
Leucine
Lysine
Average
% Standard
CRL4
21.9
7.9
10.4
18.5
7.8
11.1
Deviation
29.0
8.7
16.5
15.2
10.2
12.1
9.1
14.6
8.8
18.1
12.2
15.5
1.2
0.6
4.4
3.1
14.8
0.4
0.2
14.6
1.8
1.6
4.4
2.6
9.2
2.7
1.7
15.2
1.7
0.9
3.7
2.6
11.2
2.5
1.1
7.0
33.2
80.1
22.8
16.2
22.0
58.0
59.8
MoIe%
MoIe%
Mole%
Literature
11.3
16.3
10.4
MoIe%
Mole%
Mole%
Mole%
MoIe%
Mole%
Mole%
CRL1
23.8
8.6
8.8
CRL2
16.1
6.9
12.2
CRL3
12.1
7.9
12.8
16.0
11.8
13.1
8.2
18.0
7.8
14.8
2.0
0.8
3.9
2.2
3.2
4.0
1.4
14.1
. 2.5
1.3
2.8
2.4
10.0
3.4
1.1
16.6
1.4
0.0
3.1
2.1
10.9
3.4
1.5
Average
% Standard
Table 2.5 The mole percent distribution of amino acids in HMW DOM.
Literature, refers to the data published by McCarthy et al. (1996). Despite large variations in the
distribution of amino acids for the CRL series, total yields ranged from 15-19 % (of the total nitrogen in the
sample).
Using the fact that arganine and lysine contain 3 and 1 moles of nitrogen, respectively,
that do not participate in peptide bonds, the relative percent of free amines in a protein
containing the relative distribution of amino acids shown in Table 2.4, can be estimated
to be approximately 12% of the total nitrogen. This calculation assumes that all the
amino acids are present in one protein, thus an additional mole of free amine nitrogen
(from a terminal amino acid) is also included in this percentage. These amino acids
account for approximately 20% of the total nitrogen in the sample and thus the free amine
contribution to the 15NNMR spectrum shown in Figure 2.10A is approximately 2.4 %of
the total nitrogen in the sample, leaving approximately 5-6% of the free amine
unaccounted for. These remaining free amines may come from either amino sugars (5% N by hydrolysis in section 2.3.2.3) or other proteins.
2.4
Discussion
The goal of this chapter was to determine the chemical characteristics HMW
DOM isolated from the ocean by ultrafiltration. Furthermore, by isolating and
characterizing HMW DOM for different parts of the ocean, this study attempted to
determine the overall variation in the chemical structure of HMW DOM.
In accordance with other published studies (Benner et al., 1992; McCarthy et al.,
1996; Sakugawa and Handa, 1985b; Vernonclark et al., 1995), using 1HNMR
spectroscopy this study found that a major fraction of HMW DOM in surface waters (6070% of the total carbon), and a significant fraction of HMW DOM isolated from deep
waters (30-40% of the total carbon) consisted of identifiable (Table 2.2). The results in
Table 2.1 indicated that on average, 27 % of the total DOC was isolated by ultrafiltration.
Thus, in the surface and deep ocean, the major biochemicals identified above respectively
account for 16-19 and 8-11 % of the total DOC.
The major identifiable biochemical in both surface and deep waters was
carbohydrate, consisting of a fixed ratio of neutral monosaccharides (16±3 % rhamnose,
16±2 % fucose, 8+3 % arabinose, 13±2 % xylose, 14±2 % mannose, 20±3 % galactose,
and 14±4 % glucose). Due to the constant mole percent distribution of monosaccharides
in both surface and deep water HMW DOM isolated from these diverse locations, the
monosaccharides are postulated to be present together in a single class of
polysaccharides. In addition, the relatively constant carbohydrate: acetate ratio in all
these samples implies that acetate is also present in these polysaccharides, linked to Oacetylated or N-acetylated sugars. These polysaccharides, containing acetate and the
relative distribution ofmonosaccharides that were identified above, are referred to as
acylated polysaccharides or APS, for the purpose of this study.
Hydrolysis of HMW DOM and subsequent monomer analysis showed
quantitative differences in the yield of major biochemicals compared with NMR data.
Neutral monosaccharides as analyzed by alditol acetates accounted for 15-20% of the
total carbon in surface waters (compared with the 50-60% yield expected by NMR
spectroscopy). Amino sugars and uronic acids accounted for approximately 5-10% of the
total carbon (amino sugar yields by both GC and HPLC are comparable at 5%), thus
increasing the carbohydrate carbon yield to approximately 25%. Lipid analysis showed
only acetate as the major lipid in HMW DOM, accounting for approximately 8% of the
total carbon. In addition amino acid analysis showed that 4% of the carbon in HMW
DOM could be accounted for as hydrolyzable amino acids. When these latter carbon
yields are added to the carbohydrate carbon yield of 25%, it is clear that hydrolysis
products can account for only 37% of the carbon in HMW DOM.
The same calculation can be done for the nitrogen in HMW DOM. The Ophthalaldehyde derivatization performed by Dr. Silvio Pantoja (as in section 2.28)
showed that 20 and 5% of the nitrogen in hydrolyzable HMW DOM were amino acids
and amino sugars, respectively.
The ultrafiltration data given in Table 2.1, together with published data (Benner et
al., 1992; Guo et al., 1997; McCarthy et al., 1996), suggest that there is no systematic
depth related difference in the fraction of DOM isolated by ultrafiltration. Thus, since
DOM concentrations in the surface ocean are double those in the deep ocean (and DOM
is recycled over several ocean mixing cycles), surface waters can be assumed to be a
mixture of 50 % deep water HMW DOM (refractory) and 50 % surface water HMW
DOM (newly added). Assuming that the compounds labile to acid hydrolysis (APS for
example) are derived solely from the newly added surface water component, then 7075% of this new carbon can be accounted for by the hydrolysis products. This
calculation is illustrated schematically in Figure 2.23.
100%
50%
0%
New Surface DOM
0%
Deep, Refractory DOM
37%
Hydrolyzable
Acetate and
Monosaccharides
Amino Acids
50%
100%
Figure 2.23 The total amount of carbon in surface water HMW DOM derived from refractory deep
ocean HMW DOM, and HMW DOM newly added at the surface (A) and the relative amount of
carbon accounted for by the hydrolysis products (B), assumed to be derived from new HMW DOM
only.
As discussed earlier, the hydrolysis products identified in Figure 2.23 accounted
for only 25% of the total nitrogen in HMW DOM. In Section 2.3.1 it was shown that
there was no depth related change in the C/N ratio of HMW DOM. In both surface and
deep waters, the average C/N ratio was shown to be 15, indicating that both the newly
added component of HMW DOM in the surface ocean, and the upwelled refractory
HMW DOM component must have a C/N ratio close to 15. In order to balance the
carbon and nitrogen budget for surface waters, the remaining 25% of the new HMW
DOM (Figure 2.23B) in surface waters (and thus 13% of the total) must have a C/N ratio
< 6. Common biochemicals with C/N ratios similar to this value are N-acetylated (8)
sugars and proteins (< 5). In order to close the carbon budget for the new HMW DOM
component added to surface waters, more compositional information is needed for the
nitrogen containing compounds in HMW DOM.
The 15NNMR data presented in section 2.3.2.3, showed that 92% of the nitrogen
in surface HMW DOM is present as amide nitrogen derived from N-acetylated sugars or
proteins. Mild hydrolysis of this HMW DOM converts approximately 70% of the amide
nitrogen to amine nitrogen. The remaining amide (33% of the total nitrogen) must be
derived from non-hydrolyzable proteins and not N-acetylated sugars, because the NMR
data suggest complete hydrolysis of the acetate peak.
Several studies have shown the
presence of refractory, and hydrolysis resistant proteins both in the water column and in
sediments (Keil and Kirchman, 1991; Knicker et al., 1996; Tanoue et al., 1995),
indicating that unhydrolyzable proteins may be the nitrogen component upwelled from
deep waters. As mentioned before, the hydrolysis products showed the presence of
amino acids, accounting for 20% of total nitrogen in the sample. Thus 53% of the total
nitrogen in the samples can be accounted for by hydrolyzable and non-hydrolyzable
proteins. Thus the remaining 46 % of the nitrogen in hydrolyzed HMW DOM is amine
nitrogen. The remaining amine compounds are derived from the hydrolysis of compounds
containing amide bonds (not proteins, unless the amino acids generated upon hydrolysis
are not identifiable by HPLC).
If 46 % of the nitrogen in HMW DOM (present as amines after hydrolysis) is
derived from amino sugars, then given their C/N ratio of 6, amino sugars could account
for approximately 20% of the carbon in HMW DOM (average C/N ratio of 15). As
shown in Figure 2.23, the amount of carbon needed to close the carbon budget for the
new component of HMW DOM added in surface waters, is 13 %. Thus, amino sugars
could easily account for the missing new HMW DOM component in surface waters.
Furthermore, combining the amino sugar carbon with the amount of carbon derived from
hydrolyzable monosaccharides increases the total carbohydrate yield to 45 %. This value
is similar to the amount of carbohydrates estimated by both 'H and ' 3 CNMR. The
amount of acetylated amino sugars in HMW DOM can also be calculated if all the acetate
(8%) in HMW DOM is assumed to be present within these sugars. This calculation
indicates that 30% of the total HMW DOM carbon should be derived from amino sugars,
and is higher than the '5 N NMR estimate. However, the NMR data ('H, ' 3 C and '5 N), the
constant ratio of carbohydrate to acetate, and the nitrogen and carbon budget in HMW
DOM, are all consistent with the presence of N-acetylated sugars, (Figure 2.24). Other
authors have also proposed that amino sugars are abundant in HMW DOM (Boon et al.,
1999; McCarthy et al., 1998)). McCarthy et al. (1998) proposed that the major fraction
of nitrogen in HMW DOM is derived from peptidoglycans, a resistant bacterial
membrane component. These compounds contain two acetylated amino sugars, N-acetyl
glucosamine and muramic acid, in equal abundance. However, repeated attempts to
hydrolyze the amino sugars in our study, using the different hydrolysis methods
discussed in section 2.3.5 were unsuccessful even though standard glucosamine polymers
(Chitin) were completely hydrolyzed by three of the four methods. Thus no direct
identification of amino sugars was possible to confirm the data shown in Figure 2.24.
However, hydrolysis of whole bacterial cells with 4 M HCl at 100 0 C for 19 h, yielded
both muramic acid and N-acetyl glucosamine, indicating that peptidoglycans are
hydrolyzable. No muramic acid was detected upon hydrolysis of HMW DOM indicating
that if peptidoglycans are a significant fraction of HMW DOM, they are resistant to
hydrolysis and must therefore be modified compared to intact bacterial cell walls.
Furthermore, the relatively constant relationship between neutral monosaccharide yields
and acetate concentrations indicate that the N-acetylated sugars are closely coupled with
the neutral monosaccharides. This relationship would not be expected if amino sugars
and neutral monosaccharides were present within distinct biochemicals, as would be the
case if amino sugars were derived from peptidoglycans. Thus the data from this chapter
suggests that peptidoglycans are not a significant source of the N-acetylated sugars in
HMW DOM.
The preceding discussion determined that acylated polysaccharides (APS) isolated
from the surface ocean consist of seven neutral monosaccharides, and N-acetylated amino
sugars. A hypothetical structure for APS based on this data is shown in Figure 2.25.
Since acid hydrolysis gave high neutral monosaccharide recoveries but poor amino sugar
recoveries, this figure shows the amino sugars bonded to each other within APS.
New Surface DOM
Deep, Refractory DOM
50%
37%
0%
100%
II!.''! llil iilll llilllllll
1111
Unhydrolyzable Amino Sugars
Hydrolyzable
Monosaccharides
Acetate and
Amino Acids
Figure 2.24 The total carbon in surface water HMW DOM accounted for by the biochemicals
identified in this study. Also refer to Figure 2.23
O
0-O
O
0
k0
Ny
O
Figure 2.25 Hypothetical structure for APS, containing neutral monosaccharides and N-acetylated
sugars.
Both the amount of monosaccharide carbon in HMW DOM and the acetate and
carbohydrate resonances in the NMR spectra decrease with depth indicating that APS
decreases with depth. This is consistent with the idea that the source of APS resides in
surface waters and APS is thus the new component of HWM DOM added to upwelled
deep water in the surface ocean. If the new nitrogen in surface waters is accounted for by
amino sugars, then most of the nitrogen in the deep ocean must be protein (from the
15
NNMR data), consistent with the presence of non-hydrolyzable (likely refractory)
proteins. The elevated concentration of proteins in deep waters relative to surface waters
is also consistent with the 'HNMR spectra of HMW DOM isolated from deep waters,
which showed a relative increase in the area of the baseline between 1 and 3.0 p.p.m
(Figure 2.7 B). Resonances in this region are abundant in proteins (Figure 2.9).
However, since the C/N ratio of HMW DOM in the deep ocean is 15, there must be a
significant fraction of compounds other than proteins (C/N = 5) in deep ocean HMW
DOM that remain to be identified. A 15NNMR experiment similar to that conducted in
section 2.3.2.3 should be performed on HMW DOM isolated from the deep ocean to
confirm that proteins are the major nitrogen component.
2.5
Conclusions
The data presented in this chapter show that HMW DOM isolated from different
parts of the ocean is chemically related. The chemical analyses performed in this chapter
show that HMW DOM consists of identifiable biochemicals, namely carbohydrates,
proteins and lipids. The carbohydrate component referred to as APS has been shown to
consist of seven neutral monosaccharides, rhamnose, fucose, arabinose, xylose, mannose,
galactose and glucose. In addition to the neutral monosaccharides, the presence of
acetate and the results from 1SNNMR infer that N-acetylated sugars are also a major
component (-50% of the total carbohydrates) of APS. However, no direct detection of
these amino sugars was possible indicating that they are resistant to hydrolysis. The data
also indicate that APS is 50% of the total carbon in HMW DOM isolated from surface
waters and 15-20% of the carbon isolated from deep waters (calculated using the relative
amount of acetate present in the deep ocean). Twenty percent of the nitrogen in surface
waters is recoverable as proteins. Nitrogen NMR further indicates that the major nitrogen
containing component in the deep ocean are hydrolysis resistant proteins, while in surface
waters amino sugars are an equally dominant fraction of the nitrogen.
2.6
References
Aluwihare L. I., Repeta D. J., and Chen R. F. (1997) A major biopolymeric component to
dissolved organic carbon in surface seawater. Nature 387, 166-169.
Benner R., Pakulski J. D., McCarthy M., Hedges J. I., and Hatcher P. G. (1992) Bulk
chemical characteristics of dissolved organic matter in the ocean. Science 255,
1561-1564.
Beardsley R. C., Boicourt W. C., and Hausen, D. (1976) Physical Oceanography of the
Middle Atlantic Bight. Limnology and Oceanography,Special Symposium 2, 2034.
Benton F. L. and Dillon T. E. (1942) Journalof the American Chemical Society 64, 1128.
Boon J. J., Klap V. A., and Eglinton T. I. (1999) Molecular characterization of
microgram amounts of oceanic colloidal organic matter by Direct Temperature
resolved ammonia Chemical Ionization Mass Spectrometry. Organic
Geochemistry submitted.
Borch N. H. and Kirchman D. L. (1997) Concentration and composition of dissolved
combined neutral sugars (polysaccharides) in seawater determined by HPLCPAD. Marine Chemistry 57, 85-95.
Buesseler K. O., Bauer J. E., Chen R. F., Eglantine T. I., Gustafsson O., Landing W.,
Mopped K., Moran S. B., Santschi P. H., VernonClark R., and Wells M. L. (1996)
An intercomparison of cross-flow filtration techniques used for sampling marine
colloids: Overview and organic carbon results. Marine Chemistry 55, 1-31.
Burney C. M., Johnson K. M., Lavoie D. M., and Sieburth J. M. (1979) Dissolved
carbohydrate and microbial ATP in the North Atlantic: concentrations and
interactions. Deep Sea Research 26, 1267-1290.
Carlson D. J. and Mayer L. M. (1985) Molecular weight distribution of dissolved organic
materials in seawater as determined by ultrafiltration: A re-examination. Marine
Chemistry 16, 155-171.
Degens E. T. (1970) Molecular nature of organic nitrogen compounds in seawater and
recent marine sediments. In: D. W. Hood (ed.), Organic Matter in Natural
Waters. Univ. of Alaska., pp. 77-106.
Gagosian R. B. and Lee C. (1981) Processes controlling the distribution of biogenic
organic compounds in seawater. In: E. K. Duursma and R. Dawson (ed.).
Elservier, Amsterdam. Marine Organic Chemistry, pp. 91-123.
Gagosian R. B. and Stuermer D. H. (1977) The cycling of biogenic compounds and their
diagenetically transformed products in seawater. Marine Chemistry 5, 605-632.
Guo L., Coleman C. H. J., and Santschi P. H. (1994) The distribution of colloidal and
dissolved organic carbon in the Gulf of Mexico. Marine Chemistry 45, 105-119.
Guo L. and Santschi P. H. (1996) A critical evaluation of the cross-flow ultrafiltration
technique for sampling colloidal organic carbon in seawater. Marine Chemistry
55, 113-1127.
Guo L., Santschi P. H., Cifuentes L. A., Trumbore S., and Southon J. (1997) Cycling of
high molecular weight dissolved organic matter in the Middle Atlantic Bight as
revealed by carbon isotopic (13C and 14C) signatures. Limnology and
Oceanography.
Gust D., Moon R. B., and Roberts J. D. (1975) Applications of natural abundance
nitrogen- 15 nuclear magnetic resonance to large biochemically important
molecules. Proceedingsof the NationalAcademy of Science 72(12), 4696-4700.
100
Gustafsson O., Buesseler K. 0., and Gschwend P. M. (1996) On the integrity of the
cross-flow filtration for collecting marine organic colloids. Marine Chemistry 55.
Hakomori S.-I. (1964) A rapid permethylation of glycolipid, and polysaccharide
catalyzed by methylsulfinyl carbanion in dimethyl sulfoxide. Journalof
Biochemistry 55(2), 205-208.
Hedges J. I. (1992) Global Biogeochemical cycles: progress and problems. Marine
Chemistry 39, 67-93.
Ittekot V., Brockman U., Michaelis W., and Degens E. T. (1981) Dissolved free and
combined carbohydrates during a phytoplankton bloom in the Northern North
Sea. Marine Ecology ProgressSeries 4, 299-305.
Keil R. G. and Kirchman D. L. (1991) Dissolved combined amino acids in marine waters
as determined by a vapor-phase hydrolysis method. Marine Chemistry 33, 243259.
Klap V. A. (1997) Biogeochemical aspects of salt marsh exchange processes in the SW
Netherlands. Ph. D. Thesis, Netherlands Institute ofEcology.
Knicker H. A., Scaroni A. W., and Hatcher P. G. (1996)
13 C
and
15 N
NMR
spectroscopic investigation on the formation of fossil algal residues. Organic
Geochemistry 24, 661-669.
101
McCarthy M., Hedges J. I., and Benner R. (1996) Major biochemical composition of
dissolved high molecular weight organic matter in seawater. Marine Chemistry
55, 282-297.
McCarthy M. D., Pratum T., Hedges J. I., and Benner R. (1998) Chemical composition of
dissolved organic nitrogen in the ocean. Nature 390, 152-153.
Mopper K. (1977) Sugars and uronic acids in sediment and water from the Black Sea and
North Sea with emphasis on analytical techniques. Marine Chemistry 5, 585-603.
Pantoja S., Lee C., and Lee C. (1999) Molecular weight distribution of proteinaceous
material in Long Island Sound sediments. Limnology and Oceanography(in
press).
Peltzer E. T. and Brewer P. G. (1993) Some practical aspects of measuring DOC sampling artifacts and analytical problems with marine samples. Marine
Chemistry 41, 243-252.
Sakugawa H. and Handa N. (1983) Chemical studies of dissolved carbohydrates in
seawater. Part 1. The concentration and separation of dissolved carbohydrates.
Journalof Oceanographyof the Society of Japan39, 279-288.
102
Sakugawa H. and Handa N. (1985) Isolation and chemical characterization of dissolved
and particulate polysaccharides in Mikawa Bay. Geochimica et Cosmochimica
Acta 49, 1185-1193.
Santschi P. H., Guo L., Baskaran M., Trumbore S., Southon J., Bianchi T. S., Honeyman
B., and Cifuentes L. (1995) Isotopic evidence for the contemporary origin of high
molecular weight organic matter in oceanic environments. Geochimica et
CosmochimicaActa 59(3), 625-631.
Sidorczyk Z., Zahringer U., and Rietschel E. T. (1983) Chemical structure of the lipid A
component of the lipopolysaccharide from a Proteus mirabilisRe-mutant.
EuropeanJournalofBiochemistry 137, 15-22.
Skoog A. and Benner R. (1997) Aldoses in various size fractions of marine organic
matter: implications for carbon cycling. Limnology and Oceanography42, 18031813.
Tanoue E., S. N., Kamo M., and Tsugita A. (1995) Bacterial Membranes: A possible
source of a major dissolved protein in seawater. Geochimica et Cosmochimica
Acta 59(12), 2643-2648.
Vernonclark R., Goldberg E. D., and Bertine K. E. (1995) Organic and inorganic
characterization of marine colloids. Chemisty and Ecology 11, 69-83.
103
York W. S., Darvill A. G., McNeil M., Stevenson T. T., and Albersheim P. (1985)
Isolation and characterization of plant cell walls and cell wall components.
Methods in Enzymology 118(3-40).
104
2.7
Appendix
2.7.1
Nuclear Magnetic Resonance Spectroscopy Conditions
3
'HNMR
CNMR
AU PROG:
PRESAT.AUR
SF
SY
01
SI
TD
SW
HZ/PT
PW
RD
AQ
RG
NS
TE
FW
02
DP
LB
GB
CX
CY
Fl
F2
HZ/CM
PPM/CM
SR
SF
SY
01
SI
TD
SW
HZ/PT
300.134
210.0
5533.103
32768
16384
4201.681
.256
6.0
0.0
1.950
80
12108
297
5300
5533.103
28L
0.0
0.0
19.0
10.0
9.000P
0.001P
142.168
.474
4092.46
105
75.469
75.0
6271.031
1024
1024
18518.519
36.169
PW
RD
AQ
RG
NS
TE
4.0
2.0
0.028
320
271325
297
FW
02
DP
P9
23200
4350.000
16H
100
LB
GB
CX
CY
Fl
F2
HZ/CM
PPM/CM
SR
100
0.0
19.0
10.0
225.205P
-19.000P
972.755
12.890
-1465.63
2.7.2
Linkage Data for Seawater
1
2
3
4
5
Sugars
Linkages
Vincent Klap
CRL
Hawaii
T
42.8
32.9
2
1,2
6.47 j
11.4
3
1,3
13.4
18.2
4
1,2,3
5
1,3,4
6
1,2,4
7
1,2,3,4
1
Rhamnose
Fucose
6.3
6.7
6.8
13.6
6
13.4
9.2
26.1
2
I
g9
69
T
6.11
25
57.9
1 0
1,2
13.6
18.4
5.3
11
1,
3
12.7
12.6
6.3
1 2
1,4
13
1,2,3
14
1,2,4
12.6
5.3
15
1,3,4
6.11
10.9
5.3
16
1,2,3,4
33.31
15.5
20
24.7
22.4
46.2
13.2
23.9
13.2 i
24.6
17
1 8 Arabinose
19
T
;1,2
12.3
20
1,3
21
1,4
6.9
22
i 1,2,4
9 .9
23
i
51.3
47j
1,2,3,4
i
24 _
13.8
25 Xylose
T
26
1,3
27
1,4
28
1,3,4
31.2i
29
11,2,4
21.5
1,2,3,4
26.8
5.3
i
76.71
35.8
5.4
9.7
22j
01
31
106
0
11.1
0
28
i
1
2
3
4
5
Sugars
Linkages
Vincent Klap
CRL
Hawaii
6.64
32 Mannose
T
33
1, 2
34
1,3
35
1,6
8.07
36
1,2,3
54.9
37
1,2,4
8.5
38
1,2,3,6
39
1,2,4,6
40
1,2,3,4,6
23.4
10
32.3
12.2
6
7
38.6
32
.38.6
21.1
41
43
1,2,3,6
44
1,3,4,6
45
1,2,3,4,6
35
0
42.9
100
28.7
57
13.4
16.6
46
47 Galactose
T
48
1,2
6.11
49
1,3
5.27
10.7
5.8
50
1,6
37.7
5.7
0
51
1,3,4
16.7
9.5
16
52
1,3,6
5.7
0
53
1,4,6
6.2
6.4
54
1,2,3,6
30.9
25.4
55
1,2,4,6
17.9
56
1,2,3,4,6
31.9
107
12.2
2.73
Retention Times for Partially Methylated Alditol Acetates
RRT (relative
to mvo-Inositolhexaacetate)
As of: 22 Mar 96
Comoound
terminal-RibD
terminal-Rhao
terminai-Araf
terminai-Fuc
terminal-AraD
terminal-Xvio
2-linked-Ribf
2-linked-Rhao
2-linked-Araf
3-linKea-Ribf
3-hinKec-Rhao
terminal-Mano
4-linkea-Rhao
3-linKec-Ribo
2-linkeo-Ribo
4-linkec-Ribo terminal-Glco3-inKeo-Araf 3-;inKea-Fuce
termnal-GaiD
4-LinKea-Fuco
2-linKea-Fuco
"3-inkea-XviD
5-linKec-Araf
4-linKec-Arao 3-linKea-Arao
2-hnKec-Arao
DB1
SP2330
0.383
-0.392
0.410
0.441
0.446
0.455
0.468
0.497
0.498
0.503
0.504
0.505
0.507
0.511
0.517
0.517
0.523.
0.524
0.525
0.533
0.54Q
0.547
0.556
0.558
0.558 .
0.561
0.564
--05E
0.570 .
0.584
- 0.540.585
0.586
0.591
0.6070.6090.613
0.613
0.616
0.617
0.6200.621
0.623
0.628.
0.628 .
-
2.3-inKem-Ribf
3.4-lnKea-Rhao -2-linkea-Xvio
__
4-linkea-Xvlo
3.4-linkea-Fuc
.
2.3-linKea-Araf
2.3-hnkea-Rhao
3-inKec-Gico
2.4-linKea-Rhao
3-linkea-Mano
2-linkea-Mano
2.3.4-hnKea-Rhao
2-linkea-Gtco 2.3-linked-Fuco
2.4-linkea-Ribo
2. 4 -linkea-Fuco
2.3-lnkea-Ribo
3.4-linked-Ribo
108
--
-
--
-
As of 22 Mar 96
Comoound
3-linked-GaiD
4-inked-ManD
- 2.3.4-4inked-Fuco
2-linked-GalD
6-inked-Mano
3,5-linked-Araf
3,4-linked-Arao
6-inked-GIco
4-inkeo-Galo
2.3-linked-AraD
4-inkea-Glco
2.5-linkea-Araf
2.4-inKe-Arao
2.3.4-linked-Ribc
2.3-linked-Mano
3.4-inked-Xvio
2.3-linkea-Xvlp
2.4-linkea-XvicD
3.4-inked-Mano
6-linkec-GalD --- 3.4-linkea-Galo
2,3.5-linked-Araf
2.3.4-iinked-Ara
= 2,3-iinkea-GalD
3.4-linKed-GicD -2.3-linkea-GIco
.2.4-linkea-Mano
2.4-linkea-Gico
2.3,4-linKed-Mano
2.4-linked-Galt
4.6-linKed-Mano
2.3.4-linked-Gala
3.6-linked-Gico
3.6-linked-Mano
2.6-linked-Mano
4.6-linkea-Gico
2.6-linked-GicD
2.3.4-linked-XvlI
4.6-linkea-Galo
2.3,4-linked-Gico
3.6-linked-GalD
2.6-linked-GalD
3.4,6-linked-Manp
3,4.6-linked-Galo
3.4.6-linked-Glco
RRT (relative
to mvo-lnositolhexaacetate)
SP2330
0.628
0.630
0.630
0.648
0.649
0.651
0.651
0.652
0.654
0.662
0.666
0.667 _
0.6670.690
0.691
0.691 0.691
0.691
0.6950.700
0.7100.715
0.715
0.716 ,
0.722 0.731
0.736
0.742
0.742
0.743
0.749
0.7640.764
0.770
0.778,
0.782
0.783
0.786.
0.790
0.792
0.797
0.815
0.815
0.841
0.849
109
DB1
As of 22 Mar 96
Comoound
2.3.6-linked-Mano
2.4,6-linked-Mano
2.4,6-linked-Gicp
2,3.4.6-linked-Manp
-2.3.6-linked-Galp
2.4,6-linked-Galp
2.3.4.6-linked-Galo .
2.3,6-linked-GIcp c2.3.4.6-linked-Gico
terminal-GicoNAc
terminal-GaloNAc
terminal-MannNAc
4-4inkea-GIcoNAc
4-inkeo-GalNAc
4-linked-ManDNAc
3-linked-ManoNAc
3-linked-Glco NAc
3-inked-GaloNAc
6-linkea-GlcNAc
3.4-linKed-ManpNAc
3,4-linked-GalpNAc
3.4-linked-GlcoNAc
6-linkea-ManoNAc
6-linKeo-GaloNAc
4.6-linkea-GlcoNAc
4.6-linked-ManoNAc
4.6-linkec-GaloNAc
3.6-linkec-ManoNAc
3.4.6-linked-ManoNAc
3.6-linked-GIcoNAc
3.6-linked-GalpoNAc
3.4.6-linked-GlcoNAc
3.4 6-linkeo-GaloNAc
RRT (relative
to myo-inositoihexaacetate)
SP2330
DB
0.871
0.873
0.878
0.881
0.905
0.905
0.915
0.916
0.953
0.965
1.022
1.024
1.071
1.091
1.103
1.122
1.132
1.146
1.150
1.152
1.176
1.177
1.197
1.198
1.239
1.261
1.273
1.286
1.306
1.307
1.323
1.338
1.343
Temoerature Program
Initial Temo
Init Time (min)
Rate (dealmin)
Final Temo (deg C)
Final Time (min)
Rate A (deg/min)
Final Temo A (deg C)
Final Time A (mln)
110
SP2330
DB1
80.0
2.0
30.0
170.0
0.04.0
240.0
20.0
140.0
0.5
2.0
220.0
0.0
20.0
300.0
5.0
3.
A comparison of the chemical characteristics of oceanic HMW DOM and
extracellular DOM produced by marine algae.
3.1
Introduction
Dissolved organic matter (DOM) in the ocean is a by-product of marine primary
production (Carlson and Ducklow, 1995; Mague et al., 1980; Williams and Gordon,
1970). Several processes may act to transfer organic matter from plankton to seawater,
for example: direct exudation by phytoplankton (Baines and Pace, 1991; Fogg, 1966;
Fogg, 1971; Hellebust, 1965; Mague et al., 1980) and heterotrophs (Nagata and
Kirchman, 1992; Stoderegger and Herndl, 1998; Tanoue et al., 1995; Tranvik, 1994),
indirect production through "sloppy feeding" (Lampert, 1978), dissolution of fecal
pellets, marine snow (Alldredge et al., 1993) or other marine aggregates, and cell lysis
(Fuhrman, 1992). While all of these processes have been shown to occur, there is little
data that directly links the DOM released by these mechanisms to DOM which
accumulates in seawater. Thus, we do not yet understand how the process of DOM
introduction may affect its ultimate fate. For example, DOM that is directly released by
plankton may be more labile than reworked DOM introduced by grazing.
Using ' 3 C-labeled bicarbonate in a mesocosm experiment, Norrman et al. (1995)
induced a mixed algal bloom, dominated by diatoms, and stimulated the release of DOM.
Bacteria preferentially utilized labeled DOM, suggesting that newly synthesized DOM is
labile, but a fraction of the labeled DOM persisted after one year of incubation. By using
stable isotopes, Norrman et al. (1995) were able to quantify recently synthesized DOM,
and trace its production, accumulation and utilization. However, the dynamics of labeled
DOM alone do not show if the long-lived fraction of DOM was directly produced by
phytoplankton, or resulted from a reworking of the extracellular DOM by bacteria. It was
also not determined if the experimentally produced DOM was chemically related to
DOM that accumulates in seawater.
As shown in Chapter 2, a fraction of oceanic DOM retained by a lkD (nominal
molecular weight) ultrafiltration membrane (hereafter HMW DOM), is rich in acylated
polysaccharides (APS) which have a characteristic acetate/carbohydrate/lipid ratio
(Aluwihare et al., 1997). The carbohydrate fraction of APS contains at least seven major
neutral monosaccharides which occur in a relatively fixed ratio and which are linked
together at specific positions. High molecular weight DOM with a similar distribution of
neutral monosaccharides has been isolated from geographically diverse locations
(Aluwihare et al., 1997; Borch and Kirchman, 1997; McCarthy et al., 1996; Sakugawa
and Handa, 1985a; Skoog and Benner, 1997), which together with our structural data led
us to conclude that a large fraction of persistent DOM in seawater is produced by direct
algal biosynthesis.
This chapter investigates the HMW fraction of extracellular DOM produced in
laboratory cultures of three algal species, Thalasssiosiraweissflogii, a diatom, Emiliania
huxleyi sp., and Phaeocystissp., both prymnesiophytes. These algae were chosen to
represent classes of phytoplankton important in oceanic primary production, and because
previous laboratory studies have shown that the extracellular DOM produced by these
phytoplankton is rich in carbohydrates. T. weissflogii and other diatoms are abundant in
both coastal and ocean open areas; E. huxleyi is a cosmopolitan algae and a major
contributor to primary production in the open ocean, and Phaeocystissp. are mucus
producing algae important in high latitude environments. The extracellular DOM
112
isolated from each culture was studied in extensive chemical detail in order to determine
which algae could be a potential source for APS isolated from the surface ocean. Our data
show that all three species do indeed exude DOM in culture and that 20-30% of the
extracellular DOM from both T. weissflogii and E. huxleyi is spectrometrically and
chemically similar to APS isolated from surface seawater. Furthermore, degradation of
the T.weissflogii exudate shows that APS is metabolically more resistant to degradation
than other polysaccharides.
3.2
Materials and Methods
3.2.1
Phytoplankton Cultures
All algae grown for this study were obtained from the Center for Culturing
Marine Phytoplankton (CCMP), at Bigelow Laboratory, in Boothbay Harbor, Maine.
Species included, Thalassiossiraweissflogii, (CCMP 1336), Phaeocystis sp. (CCMP
628), and Emilianiahuxleyi (CCMP 373). T. weissflogii and E. huxleyi were obtained
as axenic seed cultures. The seawater used for the T. weissflogii was collected from a
depth of 1000 m, south of Nantucket Island, Massachusetts (390 49' N, 700 20' W), and
the seawater used for E. huxleyi, and Phaeocystiswas collected in Great Harbor,
Vineyard Sound, Woods Hole, Massachusetts. All seawater was prefiltered (0.2 prm) and
ultrafiltered twice to remove high molecular weight (> lkD) dissolved organic carbon
(HMW DOC). Prior to adding T. weissflogii, seawater was enriched with nutrients,
vitamins and trace metals added as f/2 concentrations according to Guillard & Ryther
(1962), with an additional enrichment of sodium selenite (NaSeO3) to a final
concentration of 10-8 M. In the case of E. huxleyi, the media was enriched with F/2 - Si
medium (Guillard and Ryther, 1962), and Phaeocystis sp. was grown in K medium
113
(Keller et al., 1987). Care was taken to keep bacterial populations low throughout the
experiment by sterilizing all culture and sampling apparatus. A flow diagram of the
experimental set up is shown in Figure 3.1. T. weissflogii and E. huxleyi were grown at
23°C, in 18 1glass carboys which were continuously aerated, and illuminated at 900
pE/m 2s, 24:0 L:D (light:dark) and 14:10 L:D respectively. The cultures were sampled
daily for particulate organic carbon (POC), dissolved organic carbon (DOC), chlorophyll
a (chl-a), nutrients and cell numbers. Phaeocystiswas grown in an 18 1 glass carboy, at
26 0 C, in an incubator on a 14/10 L:D cycle, by Dr. Steven Morton at the CCMP. The T.
weissflogii culture was harvested during late log phase growth determined by previous
culture experiments and chl-a absorbance. E. huxleyi and Phaeocystiswere also harvested
during late log phase growth, based on chl-a absorbance and cell counts, respectively.
All algae were removed by filtering (Whatman GF/F) and the exudates isolated as shown
in Figure 3.1.
Subsamples (9 ml) for counting bacterial and phytoplankton cells (E. huxleyi)
were collected daily, preserved with 1 ml of 10% glutaraldehyde and stored at 4 OC.
Samples were filtered onto blackened Nuclepore 0.2 pmn polycarbonate filters and stained
using a final concentration of 0.005 % v/v acridine orange (Hobbie et al., 1977). Five
vertical and horizontal fields were counted for each slide using a standard Zeiss
microscope with epifluorescence microscopy. The microscope was fitted with a bandpass
450-490 exciter, a FT510 chromatic beam splitter and a LP520 barrier filter. Bacterial
114
1
Indicates filtering through GF/F
and 0.2 urn filters
Samples for pigments, bacterial cell
numbers, DOC, POC, sugars and
nitrate
*
Seawater
>lkD DOC
<IkD
DOC
Enrich wit 1f/2, f/2-Si, or K medium
Algae
V
I
I
Samp le D aily
N
*
'
b
1L of GF/F filtered
seaw ater
A
Algal exudates
I
<1 kD DOC -
T. weissflogii
exuaare
1 Bacterial inoculation
( l tra fi l tratio n
Incubate in the dark
>1 kD
DOC
NMR
Monosaccharides
Linkage
TW3, TW4,
TW5, TW6
>1 kD
DOC
NMR
Monosaccharides
Linkage
Figure 3.1 Experimental set up for the culture experiments. Seawater was pre-filtered through 0.2
pm and 1 nm membranes. Both the Emiliania huxleyi and Phaeocystis experiments were terminated
after the algae were removed without further degradation of the exudate. TW3, TW4, TW5 and
TW6 refer to samples taken during the degradation of the ThalassiosiraWeissflogii exudate (dates
are given in Figure 3.3).
115
cell numbers were converted to carbon using a conversion factor of 20 fg C/cell (Lee and
Fuhrman, 1987).
The cultures were monitored daily for chl-a concentrations by filtering 15 - 45 ml
of the culture medium onto Whatman GF/F filters, extracting with 100 % acetone, and
measuring the absorbance at 664 nm. Conversion from absorbance to concentration was
achieved assuming a molar extinction coefficient (s) of 8x105 M-' cm-1 and a molecular
weight of 900 for chlorophyll a. POC analyses were performed according to the
acidification method of Hedges (1991). Culture samples (20 ml) were filtered onto
precombusted Whatman GF/F filters and frozen (-20 'C) until analysis (1-6 months).
Prior to analysis, filters were oven dried overnight at 60 0 C, exposed to HCl vapors in a
desiccator under vacuum (in the case ofE. huxleyi acidified with 100 pCof 2N HC1), and
oven dried again, overnight at 600 C. After acidification, the filters were subsampled for
triplicate analysis of carbon, nitrogen and hydrogen using a Fisons automated elemental
analyzer.
The filtrates from the POC samples were collected in combusted vials fitted with
teflon-lined caps, acidified with 50% H 3PO4 (5pl H3PO4/ml), and stored at room
temperature or at 40 C. The samples were analyzed for DOC by high temperature
catalytic oxidation (Peltzer and Brewer, 1993). Precision for triplicate analyses was
better than ±3.5%, and the upper limit for our system blank was 25 gpM, based on the
DOC concentration of our acidified carbon-free distilled water.
The T. weissflogii exudates isolated from the five carboys at the end of the
experiment (day 8 in Figure 3.2) were combined (80 1), filtered and subjected to
microbial degradation. The degradation was initiated by inoculating the exudate with 11
116
ofpre-filtered (Whatman GF/F) Woods Hole seawater. Following inoculation, the
exudate was incubated in the dark for 8 months to monitor compositional changes in the
DOC due to microbial degradation. Samples were taken periodically for ultrafiltration,
and DOC concentration measurements.
3.2.2
Ultrafiltration
Cultures were filtered through a pre-combusted Whatman GF/F filter and a prerinsed 0.2 pm polysulphone cartridge filter, and ultrafiltered to isolate the HMW DOM
using an Amicon DC-10L system equipped with two spiral wound polysulphone filter
membranes (1 nm, nominally >1 k Da). Filters were cleaned between samples as
described in Chapter 2, section 2.2.2. Eight to 10 1were ultrafiltered for each of the T.
weissflogii samples, and 14 1were ultrafiltered for the E- huxleyi and Phaeocystis
samples. Samples were concentrated to approximately 1-2 1 and diafiltered (10x), as in
section 2.2.2.2, using 2 1of de-ionized (Milli Q) water. Permeate from the final Milli Q
rinse no longer gave a visible precipitate when added to a solution of silver nitrate (0.5
g/50 ml). Mass balances of DOM for the ultrafiltration were 95 ± 15%, prior to
diafiltration. High molecular weight DOM concentrates were stored frozen in glass or
teflon bottles and lyophilized prior to analysis.
3.2.3
Analyses
Proton Nuclear Magnetic Resonance ('HNMR) spectroscopy and
monosaccharide, linkage and amino acid analyses were performed as described in
Chapter 2.
3.2.4
Dissolved Oxygen Measurements
Samples for dissolved oxygen were collected in replicate. Sixty ml glass bottles
(BOD bottles) were filled to the point of overflowing, capped with no headspace and
117
stored in the dark until analysis. Oxygen concentration measurements were made by Ms.
Amy Nolin at the Marine Biological Laboratory (Woods Hole, MA), using an automated
Winkler titration system designed by (Knapp et al., 1990).
3.3 Results
3.3.1
Phytoplankton Cultures
3.3.1.1 POC, DOC, and HMW DOC
Over the eight days of the experiment, chl-a and POC values in the T. weissflogii
culture rose steadily to 1.8 mg/L and 82 mg/L respectively, while total DOC values
fluctuated. DOC comprised 9% of the total organic carbon (TOC) in the system on day
8; HMW DOC comprised 27 % of DOC, and thus 2-3% of the TOC. E. huxleyi grew
slowly over the first 10 days of the experiment followed by exponential growth from days
10-22 (Figure 3.2), when the culture was harvested. At the time of harvesting (day 22)
the maximum POC value was 1.8 mM while DOC comprised 9-10% of the TOC. Also at
this time, HMW DOC was 25% of the total DOC or 2-3% of the TOC. For the
9
8
7
a
6
S
5
w
43
2
10
0
5
10
15
20
25
Days
Figure 3.2 Growth curve for the E. huxleyi culture expressed in cell numbers/mil.
118
Time
(day)
9
TW2
17
TW3
31
TW4
61
TW5
218
TW6
22
E. huxleyi
10
Phaeocystis
Sample
[DOC]
(pM C)
632
484
320
269
161
194
520
[HMWDOC]
(jM C)
171
179
63
58
12
49
111
%HMW
27
37
20
21
7
25
21
HMW Sugar C/N
(HMW)
(pM C)
14.0
48
10
39
7.2
10
5.6
3
8.85
1
9.6
17
12.0
54
Table 3.1 Ancillary data for the HMW DOM isolated from the phytoplankton cultures.
Day 0, in all cases (T. weissflogii (TW2, TW3, TW4, TW5 and TW6), E huxleyi and Phaeocystis), marks
the beginning of the phytoplankton cultures.
Phaeocystisculture, HMW DOC comprised 21% of the total DOC at the time of
harvesting. HMW DOC and total DOC concentrations in these cultures at the time of
harvesting are given in Table 3.1.
Figure 3.3 outlines the dynamics of the T. weissflogii culture in more detail. Chla (mg/1) and POC (mM) increased steadily from inoculation to harvesting, reaching a
POC/Chl-a ratio of 45 (Figure 3.3A). In contrast, the changes in the carbon to nitrogen
(C/N) ratio of the POC, shown in Figure 3.3B, are more dynamic over this time period.
Values rose from 4 to 6 during the first two days of the culture, remained approximately
constant for the next 3 days, and rose again sharply on day 5 to a value of 10, with little
change thereafter. The final C/N ratio of the POC was 10.6 (day 8) corresponding to a
PON concentration of 640 jiM. This value agrees well with the 600 tM loss of nitrate
from the medium (1217 pM starting concentration (f/2 concentrations) and 615 pM on
day 8).
119
DOC concentrations in the culture also fluctuate, as shown in Figure 3.3B. There
was no measurable change in DOC concentration until day 5, when the value rose sharply
by 560 gM to a maximum concentration of 842 tpM. This maximum was followed by a
rapid decrease of 260 pM to 582 gM between days 5 and 6, and a slower, total increase
of 53 gtM, over the course of the next three days to day 8, at which time the culture was
harvested (TW2 on Figure 3.3B).
The density of the bacterial population in the diatom culture, prior to harvesting,
was monitored in order to determine the extent of the bacterial contribution to DOM.
The amount of bacterial carbon was only a small percentage of the total DOC (< 1%)
throughout the culture and a small percentage of the total HMW DOC (< 2%) present at
the end of the culture. Coincident with the rapid decrease in DOC concentration from
day 5 to 6, bacterial cell numbers also increase. Although the increase in cell numbers
between day 5 and 6 corresponds to a three-fold increase in the amount of bacterial
carbon, this change accounts for only 3% of the loss in total DOC. If bacterial
degradation alone was responsible for this DOC loss, nearly all the DOC (97%) must
have been respired as CO 2 and not incorporated into the bacterial biomass. Bacterial
growth yields are usually in the range of 10 - 20 % (Kirchman et al., 1991; Smith et al.,
1995). As the growth yields of 3% calculated above for our culture are unusually low, it
is unlikely that bacteria are solely responsible for the loss of DOC. Other mechanisms,
such as flocculation of DOC, must have played a role in the removal of DOC from our
system. Bacteria were also present in the E. huxleyi and Phaeocystis cultures throughout
the growth period, but accounted for less than 2% of the total DOC.
120
-6
-5
0
POC
-3
3
-2
1
0
0
2
6
8
10
50
100 150 200
Days
Figure 3.3 Ancillary data from the Thalassiosiraweissflogii culture. (A) Changes in chl-a (filled
squares) and POC (filled triangles). (B) Changes in the DOC (filled squares) concentration and C/N
ratios of the POC (open triangles). Algae were removed on day 8. TW1, refers to the ultrafiltration
sample taken during early log phase, TW2 was taken during late log phase, and TW3, 4, 5 and 6 were
taken during the course of the degradation experiment after the algae had been removed. After day
9, changes in DOC concentration result from the action of microorganisms present in the incubation,
or from physical removal mechanisms.
121
3.3.1.2 'HNMR Data
Proton nuclear magnetic resonance (1HNMR) spectroscopy can be used to
determine the major biochemicals present in each of the phytoplankton exudates. The
'1HNMR spectra for HMW DOM isolated from the cultures at the time of harvesting are
shown in Figure 3.4, along with a spectrum typical for seawater (Coastal Research
Laboratory, Woods Hole, Massachusetts; Figure 3.4A). In all spectra, carbohydrate
resonances dominate, with resonances between 3.2 and 4.7 p.p.m (CHOH), 5 and 5.8
p.p.m (anomeric protons), and 1.3 p.p.m (CH 3 , from rhamnose and fucose). Resonances
at 2.0 p.p.m (CH 3COO) arise from acetate and resonances at 0.9 p.p.m (CH 3), and other
resonances at 1.3 p.p.m (CH 2) may arise from lipids. There is a close resemblance
between the spectra of seawater and the exudates from T. weissflogii and E. huxleyi
(Figure 3.4B and C). Although both exudates show the presence of carbohydrates, acetate
and lipids, the proportions of these constituents are different from seawater. The ratio of
carbohydrates:acetate:lipid carbon, calculated as in Chapter 2 (section 2.3.2.1), using
'HNMR peak areas, is 89:7:4 for the T. weissflogii exudate, and 89:6:3 for the E. huxleyi
exudate. For comparison, this ratio in seawater HMW DOM is 80-4:10±2:9-4. The
'HNMR spectrum of the Phaeocystisexudate is different (Figure 3.4 D) from the other
exudates. In this spectrum, as with the others, the carbohydrate resonances dominate
between 3.2 and 4.7 p.p.m. However, the acetate resonance at 2.0 p.p.m is smaller and the
lipid resonance at 1.3 p.p.m is larger compared to the other exudates. The ratio of
carbohydrate:acetate:lipid carbon in this sample is 90:2:5, showing the relative excess of
carbohydrates compared to seawater. For all spectra (including seawater), the
122
A. Coastal Seawater
---
B.
r
--
C5
4
Alkyl
C
PPM
T. weissflogii
...---
5
6
C.
Ac
4
PPM
3
2
3
2
1
E. huxleyi
-
__
1
3
PPM
D. Phaeocystis
6
5
4
PPM
1
Figure 3. 4 'HNMRspectra for HMW DOM isolated from seawater (Woods Hole) (A), and the
exudates of Thalassiosira weissflogii (B), Emiliania huxleyi (C), and Phaeocysis(D). Carbohydrate
resonances are between 3.2- 4.7 p.p.m (CHOH), 5-5.8 p.p.m (anomerics) and 13 p.p.m (CH3 from
deoxy sugars). Spectra also show proton resonances for acetate (2.0 p.p.m) and lipids (13 and 0.9
p.p.m). The ratios of carbohydrate:acetate:lipid reported in the text are based on the areas under
the peaks denoted C (for carbohydrates), Ac (for acetate) and Alkyl (lipids). To convert from
hydrogen to carbon we use a C/H ratio of 1 for carbohydrates, 0.67 for acetate and 0.5 for lipids.
123
carbohydrate, acetate and lipid resonances discussed above account for approximately
70% of the total carbon in each HMW DOM sample.
3.3.1.3 Monosaccharide Data
The relative distribution of neutral monosaccharides in the HMW fraction of each
of the three exudates and seawater, analyzed as alditol acetates, are listed in Table 3.2 and
shown in Figure 3.5. The seawater sample shown in Figure 3.5 is the average
monosaccharide distribution in ten APS samples. Average seawater shows equal
proportions of rhamnose, fucose, xylose, mannose and glucose (14+1%) with galactose
being slightly enriched (21%) and arabinose being slightly depleted (6%) (Figure 3.5).
While the exudates shown in Figure 3.5 also contain the same seven neutral
monosaccharides, the relative proportions of these monosaccharides differ substantially
for each sample. The relative distribution of monosaccharides in the T. weissflogii
exudate shows 4.3+1% each, of rhamnose, fucose, arabinose and glucose; 15% xylose,
37% mannose and 30% galactose. In the case of E. huxleyi, the proportions are 5%
fucose, 9±1% each of rhamnose, mannose and glucose, 12% xylose, 30% arabinose, and
24% galactose. The major monosaccharides in the Phaeocystis exudate are mannose
(30%), rhamnose (21%), arabinose (17%) and galactose (16%). Xylose, glucose and
fucose are also present, but in smaller quantities (6,4 and 2% respectively). This exudate
also contains an additional neutral monosaccharide, ribose, in a small quantity.
124
Arabinose
Arabinose Xylose
Xylose
Glucose
Mannose
Mannose Glucose
Galactose Ribose
Samples
Samples
Fucose
Rhamnose Fucose
Seawater
0.15
0.15
0.07
0.14
0.15
0.15
0.19
0
T.weissflogii 0.04
0.03
0.05
0.15
0.34
0.05
0.3
0
0.09
0.05
0.27
0.11
0.10
0.11
0.3
0
Phaeocystis 0.21
0.02
0.17
0.06
0.30
0.40
0.16
0.04
E. huxleyi
Galactose Ribose
Table 3.2 The relative distribution of monosaccharides in T. weissflogii, E. huxleyi, and Phaeocystis
at the time of harvesting.
Monosaccharides accounted for 15% (Seawater), 28% (T. weissflogii), 35+10% (E. huxleyi), and 49±9%
(Phaeocystis)of the total carbon in the HMW exudates. Seawater is the average distribution in 10 samples
from the Northwest Atlantic, an estuary in Oosterschelde (Aluwihare et al. 1997), the Sargasso Sea and the
Gulf of Mexico (McCarthy et al., 1996), and Mikawa Bay (Sakugawa and Handa, 1985b).
125
100%
90%
z RI
80%
.
70%
GA
60% -
GL
:oM
50%
aMX
40%
GA
30%
gR
20%10%
0%
)
,-
Figure 3.5 The relative distribution of monosaccharides in HMW DOM isolated from seawater, and
the exudates of Thalassiosira weissflog, Emiliania hrleyi, and Phaeocyss. R is rhamnose, F is
fucose, A is arabinose, X is xylose, M is mannose, GL is glucose, GA is galactose, and RI is ribose.
The seawater sample represents the average distribution of monosaccharides in ten seawater samples
(the standard deviation is less than :3% for each monosaccharide).
33.2
Bacterial Degradation of the T. weissflogi Exudate.
3.3.2.1 Changes in DOC
In order to observe chemical changes in the exuded DOM due to bacterial
degradation, T. weissflogii were harvested by filtering on day 8 and the filtrate inoculated
with Whatman GF/F filtered seawater. The changes in total DOC after inoculation are
also shown in Figure 3.3B (days 9-218) and listed in Table 3.1.
Three hundred tlVM C were lost from the DOC pool during the first three weeks
(days 9-31) of the incubation experiment. Approximately 50% of this DOC removal
126
occurred over the first week of the incubation (day 9 through 17, TW2 to TW3 from
Figure 3.2 and Table 3.1), during which time the concentration of HMW DOC showed no
change. Over this time period there must have been a balance between the production
and removal of DOC within the HMW fraction (note the changes in the 1HNMR spectra
(Figure 3.6)), total hydrolyzable monosaccharide concentrations and C/N ratios in HMW
DOM (Table 3.1) between these two time points). Thus, the net change in DOC
concentration reflects the reactivity of the <lkD (low molecular weight) fraction of DOC
during the first week of the degradation, and suggests some synthesis of HMW DOM
from LMW DOM by microbes. During the next two weeks (days 17-31; TW3 to TW4
from Figure 3.3 and Table 3.1) the concentration of HMW DOC decreased by 100 gM
accounting for 70% of the loss in total DOC. After the first three weeks, the removal of
DOC slows down considerably. From day 31 to 61 (TW4 to TW5, Figure 3.3 and Table
3.1), the 50 gM decrease in total DOC concentration is accompanied by little change in
the concentration of HMW DOC (but HMW monosaccharide concentration decrease over
this time period (Table 3.1)). Removal rates continue to decrease over the last five
months of the experiment (TW5 to TW6) resulting in a 100 jM decrease in the total
DOC concentration. The concentration of HMW DOC also decreases over this time
period, and can account for approximately 30% of the loss from the total DOC pool.
At day 218 of the experiment, when DOC concentrations had reached 161 jgM,
the filtrate was re-inoculated to determine if any microbially labile DOC remained in the
culture. Following the inoculation, oxygen concentration measurements yielded a total
consumption of 7 gM oxygen over a three month period, indicating that the remaining
127
organic matter was refractory to bacterial degradation (nutrient concentrations remained
high throughout this period).
3.3.2.2 'HNMR Data
Proton nuclear magnetic resonance ('HNMR) spectroscopy was used to monitor
the changes in HMW DOM as the total T. weissflogii exudate was degraded. Shown in
Figure 3.6 are spectra for the HMW exudate one week (Figure 3.6A, day 17, TW3) and
three weeks (Figure 3.6B, day 31, TW4) after the inoculation. The 'HNMR spectra of
TW5 and TW6 (not shown) are very similar to the 'HNMR spectra of TW3 and TW4.
However, between TW4 and TW5, there is a decrease in carbohydrate resonances relative
to other resonances (between 1 and 3 p.p.m) (the ratio of carbohydrate carbon/other
carbon decreases from 3 to 2). When these spectra are compared to the 1HNMR
spectrum of the T. weissflogii exudate at the time of harvesting (day 9 of the experiment)
shown in Figure 3.4B, new resonances in the region between 1.9 and 3.1 p.p.m are
apparent.
These resonances may be attributed to proteins excreted into the medium by
microorganisms, or to small microbes (< 0.2 pm) concentrated in the sample during
ultrafiltration. Protein resonances were removed by passage through a cation exchange
column (BioRad AG 50W-X8), which increases the C/N ratios of the samples from 10
and 7 (prior to passage through the column) to 14 (after passage through the cation
exchange column), consistent with removal of protein (nitrogen containing fraction). The
'HNMR spectra of these cleaned samples (Figure 3.6 insets) closely resemble the
spectrum of HMW DOM isolated when the algae were harvested (Figure 3.4 B) and the
seawater spectrum shown in Figure 3.4 A.
128
2
A
,4
6
4
5
PPM
3
2
F
I
6
5
3
-
2
T
PPM
Figure 3.6 1 HNMR spectra of two HMW DOM samples isolated during the degradation of the
Thalassiosiraweissflogii exudate, TW3 (A) and TW4 (B). Resonances between 1.9-3.1 p.p.m are
thought to arise from proteins which can be removed by passage through a cation exchange column
(insets).
129
3.3.2.3 Monosaccharide Data
The removal of polysaccharides during the incubation of the T. weissflogii
exudate was followed using the changes in monosaccharide concentrations. These data
are shown in Table 3.3. The overall changes occur in three stages. During the first stage
which spans the first three weeks of the incubation (TW2 (day 9) to TW4 (day 31)),
polysaccharide concentrations (Table 3.1) decrease by 80% accounting for 40% of the
decrease in the HMW DOC concentration. Mannose and galactose are preferentially
removed, compared to other monosaccharides, with an overall 90% decrease in mannose
concentration, and a 75% decrease in galactose concentration. At the end of the third
week, galactose, mannose and xylose are still the most abundant sugars but their
dominance is diminished, making the relative distribution of sugars more even. Between
TW4 (day 31) and TW5 (day 61), there is a decrease in the concentrations of all
monosaccharides (except arabinose). Rhamnose, xylose, mannose and galactose all
experience an 80-90% decrease in concentration, indicating removal of these sugars
during the second month of the degradation. Over the last five months of the degradation
(TW5 to TW6) there is an increase in fucose concentrations, and a decrease in arabinose
concentration.
130
Monosaccharide
Data
Sugars
TW2
%C
I
NAG
G
R
F+A
X
M
GA
TOTAL
0.8
1.7
1.2
2.3
3.9
11.3
9.2
30.4
TW2
Concentration
pM C
1.4
2.9
2.1
3.9
6.7
19.3
15.7
51.984
TW3
0.6
1
1.1
1.1
3
9.8
6.7
23.3
TW4
TW3
Concentration
TW5
TW4
Concentration %C
TW5
Concentration
pM C
pM C
pM C
1.1
1.8
2.0
2.0
5.4
17.5
12.0
41.7
0.8
1.4
1.4
1.7
2.6
3.3
5.6
16.8
0.5
0.9
0.9
1.1
1.6
2.1
3.5
10.6
0.1
1.2
0.3
2.2
0.3
0.5
0.4
5
0.06
0.70
0.17
1.28
0.17
0.29
0.23
2.9
TW6
%C
0.5
2
0.5
1.5
0.3
0.5
0.4
5.7
TW6
Concentration
pM C
0.06
0.24
0.06
0.18
0.05
0.06
0.05
0.7
Table 3.3 Absolute concentrations of monosaccharides at the time of harvesting (TW2) and at each of the time points during the incubation experiment
(TW3, TW4, TW5 and TW6) for Thalassiosiraweissflogii. % C refers to the percent of total carbon in the sample accounted for by each of the
monosaccharides.
3.3.2.4 Amino Acid Data
The 'HNMR data discussed in section 3.3.2.2 indicated the presence of protein
resonances in HMW DOM isolated from the incubation experiment. The data for the
amount of nitrogen and carbon associated with hydrolyzable proteins are shown in Table
3.4. Consistent with the 'HNMR data, all samples isolated from the incubation
Amino Acid Data
TW3
%C
0.64
Aspartic
1.01
Glutamic
6.46
Serine+Histidine
0.43
Glycine
Threonine+Arganine 2.65
0.64
Alanine
0.23
Tyrosine
0.18
Methionine
0.50
Valine
0.62
Ph-Alanine
0.41
Isoleucine
0.67
Leucine
0.12
Lysine
15
Total %
Concentration (giM) 27.0
Amino Acids
%N
2.01
2.53
17.93
2.70
8.28
2.68
0.32
0.44
1.26
0.87
0.85
1.40
0.51
42
7.5
TW4
%C
1.04
0.65
4.40
0.40
1.40
0.56
0.24
0.18
0.39
0.54
0.30
0.64
0.12
11
7
%N
2.34
1.16
8.80
1.80
3.15
1.69
0.24
0.33
0.69
0.54
0.45
0.96
0.36
22
2
TW5
%C
0.04
0.09
0.16
0.06
0.07
0.12
0.95
0.05
0.07
0.10
0.06
0.09
0.01
2
1.2
%N
0.07
0.13
0.27
0.23
0.13
0.30
0.79
0.07
0.10
0.08
0.08
0.11
0.04
2
0.2
TW6
%C
0.32
0.83
1.80
0.42
0.34
0.34
0.49
0.03
0.17
0.25
0.15
0.24
0.13
5
0.6
%N
0.90
1.87
4.50
2.37
0.96
1.29
0.61
0.08
0.37
0.31
0.28
0.45
0.50
14
0.2
Table 3.4 Concentrations of the individual amino acids in HMW DOM isolated during the
degradation of the T. weissflogii exudate (TW3, TW4, TW5 and TW6).
experiment contain proteins. Amino acid yields for TW3 and TW4 account for a large
fraction of the nitrogen in these samples (44 and 24%, respectively). Although the
fraction of carbon and nitrogen derived from amino acids increases from TW5 to TW6,
there is an overall decrease in the concentration of carbon and nitrogen associated with
these amino acids over the course of the incubation. As expected, the C/N ratio of the
132
total amino acid fraction is constant (between 3 and 4) during the incubation, except in
the case of TW5, when this ratio increases to 6. This data is consistent with the relative
distribution of amino acids which shows a marked relative increase in tyrosine at TW5
(30% of the total amino acid nitrogen, and 50% of the amino acid carbon), which has C/N
ratio of 9. The relative distribution of amino acids in HMW DOM isolated from the
incubation is shown in Figure 3.7. There is considerable variation in the relative
distribution of amino acids in these samples suggesting that the protein fraction is
dynamic.
40
O TW3
35o
30
E
E 25
0 20
I TW5
.9 15
5
ITW6
S10
50
Figure 3.7 The relative distribution of amino acids in HMW DOM isolated during the degradation of
the Thalassiosira weissflogii exudate, expressed as the percentage of each amino acid accounted for
by the culture sample. Amino acid are: aspartic acid, glutamic acid, serine and histidine, glycine,
threonine and arganine, alanine, tyrosine, methionine, valine, phenylalanine, isoleucine, leucine and
lysine.
133
3.4
Discussion
3.4.1
Phytoplankton Cultures
The major goals of this study were to determine if polysaccharides produced by
algae chemically resemble acylated polysaccharides (APS) isolated from seawater and to
monitor changes in the chemical characteristics of the exuded polysaccharides during
microbial degradation. As shown by 'HNMR, seawater APS consist mainly of
carbohydrates (80±4% of the total carbon), acetate (10±2% C) and lipid (9-h4% C).
Monosaccharide analyses have shown that APS are composed of seven major neutral
monosaccharides, rhamnose, fucose, xylose, arabinose, mannose, glucose and galactose
present in a relatively fixed ratio (Aluwihare et al. 1997).
The nuclear magnetic resonance ('HNMR) spectra shown in Figure 3.4 indicate
that the HMW DOM produced by T. weissflogii and E. huxleyi have spectral
characteristics similar to seawater APS. However, the relative areas under each of
resonances in these spectra indicate an excess of carbohydrate in extracellular HMW
DOM isolates (89:7:4 (carbohydrate:acetate:lipid carbon) for T. weissflogii and E.
huxleyi) compared to seawater (80±4:20±2:10±4). In the case of Phaeocystis,the
1HNMR
spectrum shows very little acetate (90:2:5, for carbohydrate:acetate:lipid
carbon).
When the distribution of monosaccharides in each of the exudates (Figure 3.5) is
compared to seawater, the presence of polysaccharides other than APS is apparent.
Unlike seawater, where the major polysaccharide is APS, the algal exudates are
comprised of a mixture of polysaccharides, including APS. The abundance of
polysaccharides in the exudates relative to seawater HMW DOM was indicated by the
yield of monosaccharides upon hydrolysis (Table 3.1 and Table 3.3). As shown in
134
Chapter 2, neutral monosaccharides contribute between 7-20% of the carbon in HMW
DOM (Aluwihare et al., 1997; Borch and Kirchman, 1997; McCarthy et al., 1996; Skoog
and Benner, 1997) isolated from surface seawater. At the time of harvesting, neutral
monosaccharides were 28% of the total HMW DOC in the T. weissflogii exudate,
35±10% of the E. huxleyi exudate, and 49±9% of the Phaeocystisexudate. In all cases
including seawater, the amount of carbohydrate carbon determined by 'HNMR
spectroscopy (section 3.3.2) is approximately double the alditol acetate monosaccharide
yield.
In order to investigate the similarity between seawater APS and the algal
exudates, we calculated the correlation between the relative distribution of
monosaccharides in each culture sample and seawater. The correlation coefficient
between samples is calculated by dividing the covariance in the monosaccharide
distribution of two samples (e.g. seawater vs. T. weissflogii) by the product of their
standard deviations (a). In all cases, the absolute values of the correlation coefficients are
reported. The T. weissflogii sample (at the time of harvesting) exhibits only a weak
correlation with seawater (0.34), confirming that the monosaccharide distribution in these
two samples is quite different. Two correlation coefficients were calculated for E.
huxleyi. The first (0.32) was generated using the monosaccharide distribution shown in
Figure 3.5, while the second (0.85) was generated after arabinose (the dominant sugar in
the distribution) was removed from the analysis. Without arabinose (which contributed
30% of the monosaccharides in HMW DOM), the monosaccharide distribution in the
exudate closely resembled the distribution seen for seawater. This analysis is consistent
with the conclusion that the HMW exudate from E. huxleyi mainly comprised mainly of
135
APS and an arabinose dominated polymer. For Phaeocystis,the sugar distribution cannot
be easily manipulated to increase the correlation coefficient (0.28) because several of the
monosaccharides dominate the exudate (see Figure 3.5).
While the 'HNMR data support the idea that the T. weissflogii and E. huxleyi
exudates contain APS, the monosaccharide and correlation coefficient analyses imply
that if APS is present, it is only a fraction of the total polysaccharides in each exudate. In
the case of Phaeocystis,the relative distribution of monosaccharides together with the
1HNMR
data suggests that APS (if present), is only a very small component of the total
HMW exudate.
Non APS monosaccharides in the exudates were estimated by subtracting the APS
associated monosaccharides from the total. Monosaccharides associated with APS in the
T. weissflogii and E. huxleyi exudates were calculated using the relative distribution of
monosaccharides observed in seawater, and assuming that all the rhamnose present in the
exudates was derived from APS. The data were normalized to rhamnose because it was
the least abundant monosaccharide in these exudates. The relative distribution of non
APS monosaccharides in the T. weissflogii exudate was dominated by mannose and
galactose (-40% each of the total excess). In the case of E. huxleyi, the dominant non
APS monosaccharides were arabinose (50%), and galactose (35%). The abundance of
neutral hexoses in the non APS fraction may be explained by the presence of storage
polysaccharides in the exudates. Furthermore, in contrast to the heterogeneous
monosaccharide distribution in APS, non APS polysaccharides contain mostly 2 or 3
monosaccharides. In the case of Phaeocystis,fucose was the least abundant
monosaccharide and was thus used for the calculation of APS. However, due to the low
136
concentration of fucose in this sample, the removal of APS associated monosaccharides
did not significantly change the overall distribution ofmonosaccharides in this sample.
Results from the linkage analysis for rhamnose, fucose, xylose, glucose and
galactose in the total HMW DOM isolated from seawater and the cultures at the time of
harvesting are shown in Figure 3.8 as TW2, E. huxleyi, and Phaeocystis.Most of the
linkages present in seawater are found in the culture samples, and in all cases, their
relative abundance in the culture samples fall within the range observed for seawater.
From this data it is apparent that the exudates from T. weissflogii and E. huxleyi could
contain compounds related to APS. The Phaeocystissample shows some significant
differences compared to the other two samples and compared to seawater. With the
exception of rhamnose (the second most abundant sugar), this sample has fewer linkages
per sugar, for example, 50% of arabinose present is linked at all positions; 40% of
rhamnose is terminal and 69% of xylose is linked at carbon 1,2, and 3. Fucose was only
identified linked at all positions; however, this is the least abundant sugar in the sample,
and other linkages may be present in very small quantities. The simplicity of the
Phaeocystislinkage data suggests that this exudate probably contains a much simpler
mixture of polysaccharides compared with the other two exudates.
The presence of polysaccharides in the exudates of diatoms grown in culture, has been
noted in numerous studies (Allan et al., 1972; Biersmith and Benner, 1998; Haug and
Myklestad, 1976; Hecky et al., 1973; Myklestad, 1974; Myklestad et al., 1972; Percival
et al., 1980; Smestad et al., 1975) Table 3.5). The exudates examined in Table 3.5,
contain many different monosaccharides with different linkage types, consistent with
137
Fucose
Rhamnose
100
908070-
100
9080 70605040
30
20
10
0
-10
6050
40
30
20
10
T
0
-10
;
I
:
IIIIII
I
I
te
,
S
4
I
luctns
Glucose
Xylose
00
I
100
908070605040
30
20
10
0
-10
403020100
-10
I1
tI
cf
E
Seawater
TW2
-r
Bi
It
I
I
-10 sissileO~
Cm
1
TW4
E
E. Huxleyi
[] Phaeocystis
Galactose
100
9080706050403020-
T
0-10-
1
C
e4
-
E] Seawater
O
TW2
N
TW4
(
g
It
E. Huxleyi
E]
Phaeocystis
Figure 3.8 Carbohydrate linkage data for the exudates. TW2, E. huxleyi and Phaeocystissamples
were taken when the cultures were harvested. TW4 was isolated during the Thalassiosiraweissflogii
incubation experiment. The number beneath each bar refers to the linkage site on each
monosaccharide (Le. carbon 1 and 6 as 1,6), and T refers to linkages through carbon 1 only. Only
linkages which constitute > 5% of each sugar are listed.
139
the data for T. weissflogii presented in this chapter. All the species examined produced
extracellular DOM with varying amounts of rhamnose and galactose. When fucose was
present it was a major fraction of the total monosaccharides, while xylose, arabinose,
mannose, and glucose were found in varying amounts in most of the exudates. These
data demonstrate the heterosaccharide nature of the exuded polymers in accordance with
what is observed in this study. In contrast, the extraction of whole diatom cells with
water mainly yields homopolymers such as glucans (see Table 3.5). However, cell walls
and the alkali soluble fraction of whole cells (listed in Table 3.5 under water insoluble at
pH = 7) have also been shown to contain many of the different monosaccharides
discussed above, indicating that structural polysaccharides are heterogeneous. Direct
comparison of the relative distribution of these monosaccharides indicates that for the
diatoms considered in these studies, clear compositional differences exist between
intracellular and extracellular carbohydrates. Some chemical data are also available on
extracellular DOM produced by Phaeocystis(Table 3.5). The organic mucus which
surrounds the cell has been shown to be polysaccharide (Painter, 1983), but the
monosaccharide composition of this gel-like material seems to vary among species. The
studies listed in Table 3.5 (Biersmith and Benner, 1998; Guillard and Hellebust, 1971;
Janse et al., 1996) reported the presence of rhamnose, arabinose, xylose, mannose,
galactose and glucose in the Phaeocystisexudate. In some cases, fucose, uronic acids
and small amounts of O-methylated sugars were also detected. The Phaeocystis species
grown in this study (CCMP 628) was an isolate from coastal Surinam and contained the
140
Taxon
Diatoms
Asterionella
socialis
Extracellular
R(70),F(tr),M(7),X(tr),GA
(tr), GL(tr), ?(8), ?(15)
Extract Monosaccharides
Water Insoluble (at pH=7)
R(tr),F(tr),M(22),X(tr),GA(5),
GL(tr), ?(29), ?(10)
Chaetoceros
affinis
R(35), F(39),GA(26)
R(25),F(7),RI(5),M(1 1),
X(8),GA(22),GL(3)
Chaetoceros
curvisetus
R(3),F(35),GA(10)
R(13),F(14),R(18),M(12),
X(8),GA(30),GL(5)
Chaetoceros
debilis
Chaetoceros
decipiens
Chaetoceros
socialis
Chaetoceros
species
Corethron
hystrix
Coscinodiscus
nobilis
Cyclotella
cryptica
Cyclotella
nana
R(17),F(30),M(10),X(9),
GA(29),GL(5)
R(34),F(32),M(7),X(5),
GA(17), GL(5)
R(22),F(8),RI(31 ),M(2),X(19),
GA(23), GL(4)
R(27),F(5),RI(10),M(24),X(2),
GA(10), 7(22)
R(10),F(10),RI(20),M(15),
X(12), GA(19),GL(7), ?(7)
Nitzschia
brevirostris
3.2
GL
R(3),F(31 ),RI(17),M(12),
X(12), GA(13),GL(12
R(15),F(34),M(19),X(6),
GA(tr),GL(16)
R(7),F(12),M(37),X(I 8),
GA(12), GL(13)
R(20),F(16),A(8),M(7),
X(7), GA(17),GL(tr), ?
(4), ?(21)
Allan et al., 1972
Myklestad et al.,
1972
Smestad et al., 1972
Haug and
Myklestad, 1976
Smestad et al., 1972
Haug and
Myklestad, 1976
Haug and
Myklestad, 1976
Haug and
Myklestad, 1976
Haug and
Myklestad, 1976
Haug and
Myklestad, 1976
Haug and
Myklestad, 1976
Percival et al., 1980
Hecky et al., 1973
Allan et al. 1972
R(33),F(tr),RI(8),M(10),
X(7),GA(I 4),GL( 1),
?(4), ?(13)
Cyclotella
stelligera
Nitzschia
angularis
Reference
Water Soluble (pH=7)
GL
F(0.6),A(2),M(14),X(18),
GA(22), GL(43)
Hecky et al., 1973
R(tr),F(64),M(11),X(tr),GA(4),
GL(14), ?(7)
Allan et al. 1972
R(2),F(43),RI(5),A(2),M(20),
X(4),GA( O),GL(15)
Hecky et al. 1973
Taxon
Diatoms
Skeletonema
costatum
Thalassiosira
gravida
Skeletenoma
costatum
Extracellular
Extract Monosaccharides
Water Insoluble (at pH=7)
R(12),F(0O),RI( ),A(2),
M(55),X( 11),GA(3),GL(6)
F(25),RI(27),M(12),X(12),
GA(17),GL(7)
Reference
Water Soluble (pH=7)
Handa, 1969
Haug and
Myklestad, 1976
Biersmith and
R(5), A(2),X(4),F(12),
M(14), GA(4), GL(60)
Benner 1998
PHAEOCYSTIS
sp.
Phaeocystis*
Phaeocystis
globosa
Phaeocystis
poucheti
(Woods Hole)
Phaeocystis
poucheti
(Surinam)
Phaeocystis
sp
Janse et al. 1996b
R(5),A(40),M(15),X(17),
GA(I 5),Me(8),[GL
(variable)]
GL
Janse et al. 1996a
R(I9.5),RI(3),A(5),M(27),
X(3),GA(5.5),GL(30)
Guillard and
Hellebust, 1971
R( 16),RI(8),M(64),X(1.5),
GA(2),GL(4.5)
Guillard and
Hellebust,1971
R(6),RI(2),M(16),X(13),
GA(4),GL(17),A(13),F(4)
Biersmith and
Benner, 1998
*average estimates of the data plotted in the paper - relative distribution without glucose
Table 3.5 Monosaccharide composition of polysaccharides isolated from different species of algae.
Linkage analysis was performed on (A) Chaetocerosaffinis extracellular polysaccharide: terminal rhamnose; 1,2-, 1,3-, and 3,4-rhamnose; terminal fucose; 1 3-,
1,3,4-, 1,2,3-fucose; terminal galactose and 1,3-, 1,4- galactose, (B) Chaetoceroscurvisetus extracellular polysaccharide: terminal rhamnose; 1,2-rhamnose;
terminal fucose; 1,2-,1,3-, 1,2,3-, 1,3,4-fucose; terminal galactose; 1,3-, 1,2,3-galactose, (C) Coscinodiscus nobilis extracellular polysaccharide: terminal
rhamnose; 1,2-, 1,2,3-, 1,2,4- and 1,3,4-rhamnose; terminal fucose; 1 3-, 1,2,4-,1,3,4-fucose; terminal xylose; 1,4-xylose; terminal mannose; 1,3-, 1,4-,1,6-,1,2,4,1,2,6-, 1,3,6 -mannose; terminal glucose; 1,3-, 1,4-, 1,2,4-, 1,3,6 glucose were detected.
142
complex mixture of neutral monosaccharides (including ribose) observed by other
investigators. In our samples the dominant sugars are arabinose, rhamnose, and mannose,
with fucose present only in a very small amount. Fucose has not been reported in the
literature to be abundant in Phaeocystisextracellular polysaccharides. While the
monosaccharide distributions discussed above were qualitatively similar to APS isolated
from seawater, these distributions alone do not allow a definitive comparison between the
polymers isolated from seawater and those exuded by phytoplankton in culture.
3.4.2
Degradation of the T. weissflogiiexudate
Norrman et al. (1995) demonstrated that DOM with a range of turnover times,
was produced by algae in culture. Here, we expanded on their study by following the
compositional changes within the polysaccharide fraction, including APS, as the total T.
weissflogii exudate was subjected to microbial degradation. These data can be used to
determine if the chemical composition of the refractory material persisting in culture
differs from the chemical composition of the compounds present at the time of exudation.
In addition, we directly compare the DOM present at the end of the degradation to the
DOM isolated from seawater.
To quantitatively compare HMW DOM left in the degradation experiment with
APS isolated from seawater (Figure 3.4 A), we used the carbohydrate to acetate carbon
ratio in each of these samples. This ratio varies little in seawater APS (8±2), and is well
established by our previous studies (Aluwihare et al. 1997). Although both inset spectra
in Figure 3.6 contain the same resonances, there is a difference in the relative abundance
of acetate. The carbohydrate/acetate carbon ratio at the time of harvesting (Figure 3.3 B)
was 15, and at TW3, after one week of degradation, (Figure 3.6 A, inset), was
comparable at 18. Two weeks later, at TW4, (Figure 3.6 B, inset) the ratio drops to 10,
143
not significantly different from the ratio of 8 (+2) observed for seawater. One month
later, at TW 5, this ratio was still 10. However, at TW6 (5 months later), the ratio of
carbohydrate/acetate carbon increased to 13, consistent with the changes in the
monosaccharide data discussed below. The overall decrease in this ratio is brought about
by a more rapid loss of total carbohydrate compared to the acetate containing (perhaps
APS) fraction. This indicates that the different polysaccharides within the exudate have
different removal rates, consistent with the findings ofNorrman et al. (1995), and as the
exudate is degraded, the remaining polysaccharide component resembles APS more
closely. Collectively, the NMR data indicate that a component closely resembling
seawater APS is produced by T. weissflogii in culture (Figure 3.3 B) and is present
throughout the incubation (Figure 3.5 A and B). Furthermore, carbohydrate/acetate
carbon ratios indicate that this APS component is concentrated relative to other HMW
components as the degradation proceeds.
The neutral monosaccharide data, for the different time points during the
incubation, also supports the accumulation of APS as HMW DOM is degraded. During
the degradation there is a shift in the monosaccharide distribution from one that is
dominated by the hexoses ( mannose and galactose) to one that is more evenly distributed
amongst the seven neutral sugars. Consistent with the NMR data, the monosaccharide
data indicate that a polysaccharide resembling APS exists within the algal exudate, but
that its presence is partially obscured by hexose-dominated polysaccharides.
Correlation coefficient analysis was used to determine the similarity between the
relative distribution of monosaccharides in each of the samples isolated from the
incubation and seawater. TW2 and TW3 are closely correlated with each other (0.9) but
144
only exhibit a small correlation with seawater (0.32 and 0.33, respectively). TW4 and
TW5 exhibit a stronger correlation with seawater, 0.7 and 0.8, respectively, showing that
as the exudate is degraded, the monosaccharide distribution in the HMW exudate
approaches the distribution observed for seawater. At the end of the incubation (TW6),
the correlation coefficient is lower than at the beginning of the experiment (0.15), due to
a relative increase in glucose and fucose. These sugars may be preferentially concentrated
with time and/or added (in the case of fucose) during the incubation. The low correlation
at TW6 is consistent with the increase in the carbohydrate to acetate ratio calculated from
the NMR spectrum, indicating a change in the relative abundance of APS between TW5
and TW6. As shown by the correlation coefficient analysis, there is a considerable
change in the relative distribution of monosaccharides over the course of the incubation.
Given these changes in the monosaccharide distribution, we were interested to see if the
linkage data showed similar variations. The linkage data from two time points, TW2 and
TW4, are compared in Figure 3.8. The linkage pattern of each monosaccharide falls
within the range observed for seawater. For certain sugars (for example the deoxy
hexoses and pentoses) the linkage pattern in TW4 more closely approached the average
seawater distribution, consistent with the correlation coefficient analysis. However, the
variations in the linkage patterns between these two time points are not reflective of the
much more pronounced difference in their relative monosaccharide distribution. Before
any specific conclusions can be drawn from these results, however, more data and
analyses are necessary.
An accumulation of the APS-like component relative to other polysaccharides in
the incubation experiment is suggested by the changes in the carbohydrate:acetate carbon
145
ratio calculated from the 'HNMR spectra and the relative distribution of
monosaccharides.
Given these data, we attempted to compare the change in the
concentration of total polysaccharides (as analyzed by alditol acetates), to the change in
the relative concentration of the component resembling APS, over the course of the
degradation (Figure 3.9). The APS fraction was calculated by assuming that all the
rhamnose in HMW DOM is present within APS at each of the time points sampled
during the incubation. Total APS concentrations are calculated using the seawater
monosaccharide distribution shown in Figure 3.5. The monosaccharide data were
normalized to rhamnose because it was the least abundant monosaccharide at the
beginning of the incubation and thus estimates the maximum possible loss of APS over
the course of the incubation. Both APS and total polysaccharide concentrations
decreased over the course of the incubation, however, the initial rate of degradation was
slower for the calculated APS fraction than for the total polysaccharides. Thus, the
percentage of APS in the total polysaccharide fraction increases from 31 to 68%, over the
course of the incubation. This data suggests that although APS is labile, there are other
polysaccharides within the exudate that are degraded faster than APS.
146
60
70
50
60
40
Q
g 30
8
50
- 40
0
.-
10
-30
020
50
100
150
200
250
Time (days)
-
Time vs Total Polysaccharide (±M)
Time vs APS (gM)
-
Time vs % APS
0
Figure 3.9 The change in the concentration of total neutral monosaccharides in HMW DOM and the
change in APS as a percentage of the total neutral monosaccharides (as calculated in the text) over
the course of the Thalassiosira weissflogii incubation experiment
The data in Figure 3.9 can be used to calculate the degradation rate of the total
polysaccharide fraction in HMW DOM (Table 3.6). The decrease in the total
polysaccharide concentration can be modeled as a pseudo first order removal process.
The rate constant for the degradation of the total polysaccharide fraction over the course
of the entire incubation (TW2 to TW6) is 0.055/day (+0.012). The initial rate constant
(between TW2 and TW4) is even faster: 0.07/day (r2 = 0.90). In comparison, the rate
constant for the degradation of the calculated APS fraction is 0.042 /day (+0.009) over
147
Substrate
k R[initial
Correlation(r 2)
k R[total]
Standard Error
[DOC]
-0.015
0.85
-0.015
0.007
HMWDOC
-0.023
0.73
-0.028
0.01
TP
-0.055
0.96
-0.055
0.012
TP TW2-TW4
-0.07
0.90
APS
-0.055
0.98
-0.042
0.009
APS TW2-TW4
-0.044
0.91
M
-0.08
0.96
-0.06
0.02
GA
-0.09
0.99
-0.06
0.01
-0.07
0.99
-0.06
0.01
R
-0.05
0.98
-0.04
0.01
F+A
-0.02
0.50
-0.04
0.02
GL
-0.03
0.82
-0.05
0.006
is the rate constant for the loss of each substrate over the course of the entire
incubation (in /day)
k p[initiaj] is calculated for TW2 to TW5, unless otherwise noted
kR[TOTAL]
Table 3.6 Degradation rate constants for the different organic carbon fractions in the T. weissflogii
over the course of the incubation experiment.
the entire incubation period and 0.044 IMC/day (r2 = 0.91) between TW2 and TW4,
confirming that APS is degraded more slowly than the total monosaccharide fraction.
The initial degradation rates of TP and APS will not be the same if other polysaccharides
besides APS in the TP fraction are degraded faster than APS. However, the similarity in
148
the degradation constant for both the TP and the APS over the course of the whole
experiment is expected if APS is the dominant polysaccharide in the TP fraction
remaining towards the end of the experiment (kR[total TP] not significantly different from
kR[total APS] )
Consistent with the degradation of the polysaccharide fraction, the decrease
in the concentration of individual monosaccharides can also be modeled well as a pseudo
first order removal process. The overall degradation rate constant for mannose, xylose
and galactose is 0.06 /day, while the degradation rate between TW2 and TW4 is even
faster (between 0.07 and 0.09 /day (r2 >0.96)). The overall degradation rate of rhamnose
can also be modeled as above, and is 0.04/day. The slower degradation rate of rhamnose
compared with the hexoses further confirms the fact that these monosaccharides are
initially present in different polysaccharides (e.g. hexoses are present in APS and other
polysaccharides, and rhamnose is present in APS).
Many factors can control the rates of degradation of individual polysaccharides by
bacteria (Arnosti 1994). First these molecules must be hydrolyzed extracellularly, and
second the bacteria must be able to produce enzymes that recognize the individual bonds
within the polymer. It has been speculated that structural polysaccharides are synthesized
to be particularly resistant to bacterial degradation (Thingstad and Billen, 1994). As
shown in Table 3.5 for example, several investigators have noted the heterogeneity in the
distribution of monosaccharides in diatom cell walls. This heterogeneity may impart a
resistance to bacterial degradation. Both the culture exudates and seawater APS exhibit
complex monosaccharide distributions. The ubiquity of APS and their presence at depth
in the ocean (Chapter 2 and Chapter 4) lead us to hypothesize that APS would be
149
refractory due to its complex monosaccharide distribution, consistent with the data
presented in Figure 3.9, showing the accumulation of APS in the incubation experiment.
In a recent study, Carlson et al. (1998) investigated the quality of DOM available
to bacteria in the Ross Sea polynya and the Sargasso Sea. The bloom in the Ross Sea
was dominated by an Antarctic Phaeocystis species, while the Sargasso Sea bloom was
dominated by picoplankton (with some haptophytes and diatoms also present). A
comparison between the two sites showed that the DOM produced in the Ross Sea was
readily available to bacteria, while approximately 50% of the new DOM produced in the
Sargasso Sea escaped microbial degradation, suggesting an inherent difference in the
DOM produced at these locations. This study supports our findings on the interspecies
(Phaeocystisvs. T. weissflogii and E. huxleyi) differences we observe in the chemical
composition of HMW DOM exudates, and the differences in the relative lability of
dissolved organic compounds produced during algal photosynthesis.
3.5
Conclusions
The chemical characteristics of extracellular HMW DOM from three species of
marine phytoplankton were compared to HMW DOM in seawater. Thalassiosira
weissflogii, Emilianiahuxleyi, and Phaeocystis sp., were grown in nutrient enriched
seawater that had been previously ultrafiltered to remove HMW DOM. The extracellular
HMW DOM produced in these cultures was isolated by ultrafiltration and characterized
using 1HNMR, and molecular level analyses. All species exude DOM rich in
polysaccharides, and the exudates of T. weissflogii and E. huxleyi closely resemble APS,
previously identified as a major constituent of naturally occurring marine HMW DOM.
APS may comprise up to 2-3% of the total organic carbon and 30% of the HMW DOC at
150
the time of production. Degradation of the T. weissflogii exudate alters the chemical
composition of the HMW DOM, which we attribute to differences in the reactivity of
specific polysaccharides. The component within the exudate that most resembles
seawater HMW DOM has a slower degradation rate relative to the total polysaccharide
fraction. This study indicates that APS isolated from the surface ocean can have a direct
algal source and that APS may accumulate in seawater as a result of its metabolic
resistance.
3.6
References
Allan G., Lewin J., and Johnson P. G. (1972) Marine Polymers IV Diatom
polysaccharides. Botanica Marina 15, 102-108.
Alldredge A. L., U. P., and Logan B. E. (1993) The abundance and significance of a class
of large, transparent organic particles in the ocean. Deep Sea Research 140(6),
1131-1140.
Aluwihare L. I., Repeta D. J., and Chen R. F. (1997) A major biopolymeric component to
dissolved organic carbon in surface seawater. Nature 387, 166-169.
Arnosti C. and Repeta, D. J. (1994) Rapid Bacterial degradation of polysaccharides in
anoxic marine sediment. Geochimica et CosmochimicaActa 58(12), 1639-2652.
Baines S. B. and Pace M. L. (1991) The production of dissolved organic matter by
phytoplankton and its importance to bacteria: Patterns across marine and
freshwater systems. Limnology and Oceanography36(6), 1078-1090.
Biersmith A. and Benner R. (1998) Carbohydrates in phytoplankton and freshly produced
dissolved organic matter. Marine Chemistry 63, 131-144.
Borch N. H. and Kirchman D. L. (1997) Concentration and composition of dissolved
combined neutral sugars (polysaccharides) in seawater determined by HPLCPAD. Marine Chemistry 57, 85-95.
152
Carlson C. A., Ducklow H., W., Hansell D. A., and Smith Jr. W. O. (1998) Organic
carbon partitioning during the spring phytoplankton blooms in the Ross Sea
polynya and the Sargasso Sea. Limnology and Oceanography43(3), 375-386.
Carlson C. A. and Ducklow H. W. (1995) Dissolved organic carbon in the upper ocean of
the central equatorial Pacific Ocean, 1992: Daily and finescale vertical variations.
Deep-Sea Research 42(2-3), 639-650.
Fogg G. E. (1966) The Extracellular Products of Algae. Annual Review of Oceanography
andMarineBiology 4, 195-212.
Fogg G. E. (1971) Extracellular products of algae in fresh water. Archive of
Hydrobiology 5, 1-25.
Fuhrman J. (1992) Bacterioplankton roles in the cycling of organic matter: The microbial
food web. In: Primary productivity and biogeochemical cycles in the sea.
Plenum Press, New York, 361-383.
Guillard R. R. L. and Hellebust J. A. (1971) Growth and the production of extracellular
substances by two strains of Phaeocystispouchetii.Journalof Phycology 7, 330338.
153
Guillard R. R. L. and Ryther J. H. (1962) Studies on marine planktonic diatoms I,
Cyclotella nana Hustedt and Detonula confervacae (Cleve). GrandCanadian
JournalofMicrobiology 8, 229-239.
Haug A. and Mykelstad S. (1976) Polysaccharides of marine diatoms with special
reference to Chaetocerosspecies. Marine Biology 34, 217-222.
Hecky R. E., Mopper K., Kilham P., and T. D. E. (1973) The amino acid and sugar
distribution compositions of diatom cell walls. Marine Biology 19, 323-331.
Hedges J. I. (1991) Marine Particles: Analysis and Characterization: Lignin, cutin, amino
acid and carbohydrate analyses of marine particulate organic matter. Geophysical
monograph 63, 129-137 (Under Chpt. 4).
Hellebust J. A. (1965) Excretion of some organic compounds by marine phytoplankton.
Limnology and Oceanography10, 192-206.
Hobbie J. E., Daley R. J., and Jasper S. (1977) Use of nucleopore filters for counting
bacteria by fluorescence microscopy. Applied EnvironmentalMicrobiology 33(5),
1225-1228.
154
Janse I., van Rissel M., Gottschal J. C., Lancelot C., and Gieskes W. W. C. (1996)
Carbohydrates in the North Sea during spring blooms ofPhaeocystis: a specific
fingerprint. Aquatic MicrobialEcology 10, 97-103.
Keller M. D., Selvin R. C., Claus W., and Guillard R. R. L. (1987) Media for the culture
of oceanic ultraphytoplankton. JournalofPhycology 23, 663-638.
Kirchman D. L., Suzuki Y., Garside C., and Ducklow H. W. (1991) High turnover rates
of dissolved organic carbon during a spring phytoplankton bloom. Nature 352,
612-614.
Knapp G. P., Stalcup M. C., and Stanley R. J. (1990) Automated oxygen titration and
salinity determination. Woods Hole Oceanographic Institution Technical Report.
WHOI-90-35..
Lampert W. (1978) Release of dissolved organic carbon by grazing phytoplankton.
Limnology and Oceanography23, 831-834.
Lee S. and Fuhnnan J. A. (1987) Relationships between biovloume and biomass of
naturally-derived marine bacterioplankton. Applied EnvironmentalMicrobiology
52, 1298-1303.
155
Mague T. H., Friberg E., Hughes D. J., and Morris I. (1980) Extracellular release of
carbon by marine phytoplankton a physiological approach. Limnology and
Oceanography25(2), 262-279.
McCarthy M., Hedges J. I., and Benner R. (1996) Major biochemical composition of
dissolved high molecular weight organic matter in seawater. Marine Chemistry
55, 282-297.
Myklestad S. (1974) Production of carbohydrates by marine planktonic diatoms. I.
Comparison of nine different species in culture. JournalofExperimentalMarine
Biology and Ecology(15), 261-274.
Myklestad S., Haug A., and Larsen B. (1972) Production of carbohydrates by the marine
diatom Chaetocerosaffinis var. Willei (Gran) Hustedt. II. Preliminary investigation
of the extracellular polysaccharide. JournalofExperimentalMarineBiology and
Ecology 9, 137-144.
Nagata T. and Kirchman D. L. (1992) Release of macromolecular organic complexes by
heterotrophic marine flagellates. Marine Ecology ProgressSeries 83, 233-240.
Norrman B., Zwiefel U. L., Hopkinson C. S., and B. F. (1995) Production and utilization
of dissolved organic carbon during an experimental diatom bloom. Limnology and
Oceanography40(5), 898-907.
156
Painter T. J. (1983) Algal Polysaccharides. The Polysaccharides2, 195-285.
Peltzer E. T. and Brewer P. G. (1993) Some practical aspects of measuring DOC sampling artifacts and analytical problems with marine samples. Marine
Chemistry 41, 243-252.
Percival E., Rahman M. A., and Wiegel H. (1980) Chemistry of the polysaccharides of
the diatom CoscinodiscusNobilis. Phytochemistry 19, 809-811.
Sakugawa H. and Handa N. (1985a) Chemical studies on dissolved carbohydrates in the
water samples collected from the North Pacific and Bering Sea. Oceanologica
Acta 8(2), 185-196.
Sakugawa H. and Handa N. (1985b) Isolation and chemical characterization of dissolved
and particulate polysaccharides in Mikawa Bay. Geochimica et Cosmochimica
Acta 49, 1185-1193.
Skoog A. and Benner R. (1997) Aldoses in various size fractions of marine organic
matter: implications for carbon cycling. Limnology and Oceanography42, 18031813.
157
Smestad B., Haug A., and Mykelstad S. (1975) Structural studies of the extracellular
polysaccharide produced by the diatom Chaetoceroscurvisetus Cleve. Acta
Chemica ScandinaviaB. 29, 337-340.
Smith D. C., Steward G. F., Long R. A., and Azam F. (1995) Bacterial-mediation of
carbon fluxes during a diatom bloom in a mesocosm. Deep-Sea Res. II42(1), 7597.
Stoderegger K. and Herndl G. J. (1998) Production and release of bacterial capsular
material and its subsequent utilization by marine bacterioplankton. Limnology and
Oceanography43, 877-884.
Tanoue E., S. N., Kamo M., and Tsugita A. (1995) Bacterial Membranes : A possible
source of a major dissolved protein in seawater. Geochimica et Cosmochimica
Acta 59(12), 2643-2648.
Thingstad F. and Billen G. (1994) Microbial degradation of Phaeocystis material in the
water column. JournalofMarine Systems 5, 55-65.
Tranvik L. (1994) Colloidal and dissolved organic matter excreted by a mixotrophic
flagellate during bacteriovory and autotrophy. Applied Environmental
Microbiology 60(6), 1884-1888.
158
Williams P. and Gordon L. (1970) Carbon-13:Carbon-12 ratios in dissolved and
particulate organic matter in the sea. Deep-Sea Research 17(19-27).
159
3.7 Appendix
3.7.1 Linkage date for the Cultures
1
R
1
2
3
4
5
6
Sugars
Linkages
TW2
TW4
E. huxleyi
Phaeocystis
9.4
T
12.8
13
12.2
35.5
2
1,2
3
1,3
32
25.6
21.4
4
1,2,3
8.9
6.7
6.2
5
1,3,4
10.8
9.4
5.4
6
1,2,4
13
11.7
7.7
7
1,2,3,4
21.7
24.8
10.8
5.2
7.9
39.1
15.6
28.3
8
9
F
T
10
11
1,2
1,3
0
0
12.4
12
1 2
1,4
21.5
10.3
13
1,2,3
14
1,2,4
26.1
14.7
15
1,3,4
12.8
12.8
18
29
1 6
_
1,2,3,4
8.4
1
160
41.9
8.7
18
11.8
15
100
18 A
T
19
1,2
20
1,3
21
1,4
22
1,2,4
23
1,2,3,4
38
1
27.3
10.8
13.9
24.4
16.8
10.3
51.7
100
42
52.8
24
25 X
T
25.7
23.6
26.5
8.3
5.8
26
11,3
14.1
6.6
18.9
27
1,4
22.4
13.2
29.4
28
1,3,4
5.8
23.2
7.5
29
1,2,4
18.6
23.2
7.5
1,2,3,4
13.3
30
_
j
69.1
6.5
10
31
3 2 M
T
33
1,2
34
11,3
35
j1,6
5.8
16.8
8.2
26.7
_
11.6
8.2
12
5.7
8.4
18.5
6.4
33.9
5.7
36
1,2,3
3 7
1,2,4
38
1,3,6
39
1,4,6
40
1,2,3,4
41
1,3,4,6
8.2
5.6
42
12.2
17.4
14
43
1,2,3,6
1,2,4,6
44
1,2,3,4,6
20.8
24.9
15.8
451
5.8
18.7
11.6
6.8
14.5
14.7
10.5
i__
161
4 6 GL
11,4
47
1,3,4
48
1,3,6
49
1,4,6
50
1"1,2,4,6
51
1,2,3,6
52
1,3,4,6
53
1,2,3,4,6
11.3
21.4
14
12.6
43.1
8.8
32.7
5
39.7
46.6
29.2
11.9
10.2
16.5
30.5
42.3
54
55 GA
iT
8.1!
10
56
1,2
57
11,3
11
58
1,6
12.3
6.5
9.3
19.9
59
1,2,3
60
1,3,4
12.8
7.3
23.1
22.5
61
1,3,6
6.5
13.4
6.5
62
!1,4,6
63
i1,2,3,6
64
i 1,2,4,6
7.7
65
1,3,4,6
8.7
66
-----
1,2,3,4,6
12.2
162
9.3
_
5.6
15.4
42.9
15.9
13
12.2
10.3
20.1
I
4. Radiocarbon values of individual monosaccharides isolated from oceanic HMW
DOM
4.1
Introduction
Carbon has three naturally occurring isotopes, 12C, 13C and
14C. 12C
is the most
abundant (99%) isotope, followed by 13C (1%), while only 1 in 1012 atoms in modern
carbon is present as
4/y).
14C.
While 12C and 13C are stable, 14C is radioactive (X= 1.2 x 10
Naturally occurring 14g is produced in the upper atmosphere from spallation
reactions of 14 N and rapidly combines with oxygen to form carbon dioxide (CO 2). This
carbon dioxide mixes throughout the atmosphere, dissolves in the oceans and, via
photosynthetic fixation of CO 2 to organic carbon, enters the biosphere. Aboveground
nuclear weapons testing in the 1950's and 1960's introduced large amounts of
radiocarbon into the atmosphere. Atmospheric levels of A'4 C (see section 1.3), expressed
relative to the pre 1950's value (i.e. Al4C = 0 %o), increased to approximately 1000 96o
(Broecker and Peng, 1982).
Diffusion of CO 2 from the atmosphere introduced bomb radiocarbon into the
ocean such that the A14 C value of surface dissolved inorganic carbon (DIC) reached a
maximum of approximately 200 %obetween 1973-1975. There has been a steady
decrease in both the atmospheric and oceanic reservoirs of 14C since their peak in the
1960's and 1970's, respectively. Today, surface ocean A'4C values of DIC are
approximately 70 %0. Since the decay constant for radiocarbon is approximately 10 9/
every 80 years, these atmospheric and surface ocean decreases, result from mixing
processes and not from radioactive decay. In particular, the decrease observed in the
163
surface ocean results from the more rapid vertical mixing of radiocarbon compared with
the slower diffusion rate of radiocarbon into the surface ocean from the atmosphere
(isotopic equilibration time scales are about 10-15 years).
The presence of bomb radiocarbon in the surface ocean has proven to be a useful
tracer for different ocean processes such as deep water formation and ocean circulation.
Of particular interest is its use as a tracer of recent, autochthonous organic matter. Any
organic matter produced by both autotrophic and heterotrophic organisms in the surface
ocean since the 1960's will contain post-bomb radiocarbon concentrations (A' 4C > -50
%o). The data presented in Chapters 2 and 3 suggest that polysaccharides (APS) isolated
from HMW DOM are produced in the surface ocean during photosynthesis, indicating
that APS must have a modem (post-bomb) A14 C value. The presence of APS in the deep
ocean, as revealed by the distribution of monosaccharides and the presence of acetate in
DOM, implies that surface-produced APS must be transported to the deep ocean.
Dissolved organic matter synthesized in surface waters can be delivered down to
the deep ocean by two possible mechanisms. Shown in Figure 4.1 are the two possible
scenarios for the delivery of APS, and therefore DOM, to the deep ocean. Both scenarios
show the production of DOM in the surface ocean, but differ in the mechanisms of DOM
transport to the deep ocean. Figure 4.1A emphasizes APS transport to the deep ocean by
the formation of deep water at high latitudes, while Figure 4.1B emphasizes that particle
dissolution in the deep ocean may also be an important source of DOM. This chapter is
based on the hypothesis that concentrations of 14 C in APS can be used to differentiate
between the two scenarios presented above. Since the formation of deep water is the
dominant mode of DIC
164
Depth (m)
0
Surface
200
Deep
Depth
I_
*
Surface
200
F Advected DOM
DOM
Deep
*
S
Degradation
Disaggregation
4500
Figure 4.1 Two scenarios for the delivery of surface produced DOM to the deep ocean. Scenario A
shows the delivery of DOM to the deep ocean via deep water formation. Scenario B shows that in
addition to advection, DOM may also be delivered to the deep ocean via bacterial degradation and
disaggregation of sinking POM.
165
transport to the deep ocean, there should be no difference between the radiocarbon value
of APS (corrected for isotopic fractionation during photosynthesis) and DIC, in the case
of Fig. 4.1A. The vertical transport of POM, on the other hand, occurs on much faster
time-scales compared to deep water formation. Thus, if POM delivers APS to the deep
ocean (as in Fig. 4. 1B), the radiocarbon values of APS should be more modern than DIC.
APS cannot be isolated from HMW DOM. However, it is possible to isolate the
individual monosaccharides that constitute APS and measure their A14 C values. This
chapter focuses on the development of a method to isolate monosaccharides from HMW
DOM for radiocarbon measurements in order to gain insight into the cycling of APS in
the ocean.
4.2
Materials and Methods
4.2.1
Sampling
Surface samples were collected from a depth of 5 m in the North West Atlantic
(CRL), Mid Atlantic Bight (MAB961) and Pacific Ocean (P961). Seawater was pumped
through a teflon hose using a diaphragm pump and pre-filtered to remove large particles
(Whatman GF/F) and bacteria (0.2 ipm Gelman Criticap cartridge filter). Deep water
samples were collected using Niskin bottles from the Mid Atlantic Bight (MAB966 and
MAB967), Sargasso Sea (SS972) and Pacific Ocean (P962) and pre-filtered as above.
All seawater samples (with the exception of the sample from the Sargasso Sea) were
ultrafiltered on board ship as described in section 2.2.2.1. Ultrafiltration membranes and
filters were cleaned as in section 2.2.2.1. Following ultrafiltration, the HMW DOM was
frozen (-20 and -40 0 C). Blanks for each sample were obtained as a 4 1Milli Q water
166
rinse of the ultrafiltration system immediately prior to sample processing. The Sargasso
Sea sample was stored in the dark, in teflon lined barrels, and ultrafiltered 2-3 weeks after
collection. After returning to the laboratory, frozen samples were thawed and diafiltered
to reduce salts as in section 2.2.2.2, lyophilized and stored at -20 oC until further analysis.
Seven samples were prepared for this study (Table 4.1). The samples included
two surface water samples from the North Atlantic Ocean: MAB961 from the Middle
Atlantic Bight, and CRL, collected at Woods Hole; a depth profile from the Middle
Atlantic Bight: MAB966 and MAB967 (300 and 750 m); a depth profile from the North
East Pacific Ocean: P961 and P962 (surface and 1600 m) and a deep water sample from
the Sargasso Sea: SS972 (1000 m depth in the Sargasso Sea).
Latitude Longitude Date
Sample
Depth
CRL
MAB961
MAB966
MAB967
SS972
P961
P962
Surface 41.5
Surface 36.78
300 m 36.7
750 m 36.7
1000 m 36.5
Surface 34.83
1600 m 34.83
70.5
75.87
74.58
74.58
65.46
123.15
123.15
Aug-94
Jul-96
Jul-96
Jul-96
Feb-97
Oct-96
Oct-96
[DOC] plM Liters
102
160
40
40
43
72
40
1300
800
400
400
400
350
431
HMWDOC %HMW HMWDOC2
mg C
mg C
276
18
39
100
45
19
33
22
135
163.0
20.0
26.9
14.0
38.4
28.9
Table 4.1.1 Sampling site and DOC data for the samples analyzed in this study.
HMW DOC refers to the DOC measured in the retentate following ultrafiltration. HMWDOC2 refers to
the amount of carbon recovered following both ultrafiltration and diafiltration.
167
4.2.2
Chemical Analyses
All the glassware used in this study was combusted (450 0 C) for 12 h. All tubing
and fittings used for high-performance liquid chromatography (HPLC) were teflon,
stainless steel or Upchurch PEEK (polyetheretherketone) (Upchurch Scientific, Oak
Harbor, WA). Samples were lyophilized on a system containing two water traps: one that
was a refrigerated isopropanol trap (at -30 0 C) and one containing isopropanol and dry ice
(at - 70 0C).
A flow chart of the analyses is shown in Figure 4.2. Subsamples (100- 500 9tg C)
were archived for radiocarbon analysis of total HMW DOM. The remaining samples and
blanks were hydrolyzed by adding 0.5 ml of Trifluoroacetic Acid (TFA) per 4 mg of C,
and heating in a sand bath for 2 h at 121 C.
Following hydrolysis, samples were lyophilized, re-dissolved in Milli Q water and
neutralized to between pH 7 and 8 with NH40H. Samples were then desalted using an
anion exchange resin (BioRex@ 5) in the chloride form. The ion exchange resin was
rinsed with 3x3 ml of Milli Q water and the columns were packed in 6", combusted
pasteur pipettes (6 or 8 mm I. D.). Samples were applied on to the column dissolved in
Milli Q water (10 mg of sample/ 2 ml of resin bed) and were eluted with 3 column bed
volumes. After desalting, some samples (UF961, SS972, P961 and P962) were filtered
(GF/F) to remove particles. Prior to HPLC separation, samples were lyophilized, redissolved in 1 - 3 ml of Milli Q water and centrifuged for 1-2 minutes in conical vials.
168
HMW DOM
I
2M TFA
A 4C
of Bulk
HMW DOM
Hydrolysis
Anion Exchange
Resin
I
Silver Column
(HPLC)
Hydrolyzed
Silver Column
(HPLC)
Unhydrolyzed
Amino Column
(HPLC)
Individual
Monosaccharides
Silver Column
(HPLC)
Individual
Monosaccharides
-
I.
A14C of
Monosaccharides
Figure 4.2. A flow chart of the various analytical steps involved with the separation of hydrolyzed
monosaccharides from HMW DOM. Boxes in bold denote HPLC steps. Silver column refers to the
cation exchange steps and amino column refers to the reverse phase separation.
169
4.2.3
High-Performance Liquid Chromatography (HPLC)
Instrumentation included a Waters Model 510 (Waters Corporation, Milford, MA)
HPLC pump equipped with a Shodex (Showa Denko K.K., Tokyo, Japan) RI-71
refractive index detector, and a Timberline (Alltech Associates, Inc., Deerfield, IL), TL430 column heater, programmed using a Timberline TL-50 temperature controller. A
Rheodyne 7125 Syringe Loading Sample Injector (Rheodyne, Inc., Cotati, CA) was
mounted directly onto the column heater, and was fitted with a 1000 l stainless steel
sample loop. Data acquisition was performed using HP Chemstation LC software
(Hewlett-Packard Company). The detector and injector were interfaced with the
computer through a Hewlett-Packard 35900E Analog to Digital Interface.
4.2.3.1 Ion Exchange
Following hydrolysis, the total monosaccharide fraction in each sample was
purified from unhydrolyzed material (Figure 4.2) using cation exchange chromatography
with silver as the counterion, on a sulphonated styrene-divinylbenzene gel. Two different
manufacturers were used, Supelco (SupelcogelTM Ag, cat # 5-9315, Supelco, Inc,
Bellefonte, PA) and Alltech (Benson carbohydrate, BC-100, Ag+; cat # 33802, Alltech
Associates, Inc, Deerfield, IL). The performance and the retention time of
monosaccharide standards on these two columns were identical. Column dimensions
were 30 cm x 7.8 mm I. D., and the bead size was 10 pnm. Columns were eluted with
Milli Q water at a flow rate of 0.5 ml/min, with the column heated to 800 C. Contrary to
the instructions which accompany these columns, the column should be heated to 80 oC
before connecting the flow in order to minimize large increases in back pressure. When
170
the column is disconnected for storage, the flow path should be disconnected while the
column is still at 80 'C.
The total weight of the sample injected on to the column did not exceed 1
mg/injection. In some cases (UF961, UF967, CRL, P961), samples were re-injected onto
the Ag column to further purify the monosaccharide fraction from unhydrolyzed material.
During injection and sample collection, samples were stored on ice to reduce
bacterial degradation of saccharides. Monosaccharide fractions were lyophilized
immediately after collection and prior to the second HPLC purification step.
Monosaccharides detected by HPLC were quantified by calibrating the detector using a
standard mixture of seven monosaccharides. The calibration was based on 6 injections of
increasing amounts of the standard mixture (between 5 -250 gg of each monosaccharide)
and was performed once per each sample and blank set. Twenty .l of the standard
monosaccharide mixture (20 pg of each sugar) was injected at least twice a day to insure
that there was no fluctuation in response factor over time.
Data from the six consecutive injections of a sugar standard on the Ag column
(Alltech Associates, Bodman Carbohydrate Column) showed good reproducibility in
retention time and peak area (Figure 4.3 A). Rhamnose and glucose eluted first with an
average retention time (flow rate 0.5 ml/min) of 14.24 1 0.11 minutes (F1), followed by
one peak for xylose, mannose and galactose, at 15.675 ± 0.16 minutes (F2). Fucose and
arabinose elute together at 16.907 ± 0.2 minutes (F3). For six consecutive injections of
equal monosaccharide concentration, the percent standard deviations for the areas under
each of the three peaks was between 3.4 to 5 %. All six monosaccharides examined
(rhamnose, fucose, arabinose, xylose, mannose, glucose and galactose) show similar
response factors on this column (0.0059 - 0.0081 pg/unit area) (Figure 4.3B).
A
mAU
F2
F1
1
r'
F+A
F3
X+M+Ga
R+G
/j
3ooj
00
I
i
0i
o-
/
100/
1/
0i
12
13
Sugar
R+G
X+Ga+M
F+A
14
14
166
15
Is
Ret Time % Stnd.
Dev.
0.8
14.241
15.675
1
16.907
1.2
172
17
17
Area
12316.9
20270.4
14816.6
48
1i-
% Stnd . Dev.
3.4
3.5
5
B
300
250
200
R = 0.9944 (Xylose+Mannose+Galactose)
U
E 150
=0.9991 (Glucose+Rhamn bse)
0.9991 (Glucose+Rhamn e)
= 0.9958 (Fucose+Arabinose)
100
50
0
0
5000
10000
15000
20000
25000
30000
35000
Area
Figure 43 (A) Reproducibility of six consecutive injections of a monosaccharide mixture on the Ag
column. Standard deviations (% Stnd. Dev) for retention times (Ret. Times) and peak areas are
reported as a percentage. (B) Calibrated detector response for the three different monosaccharide
fractions (R2 for each linear fit to the data are reported on the figure).
4.2.3.2 Normal Phase Chromatography
Separation of individual monosaccharides was achieved using a silica column,
with an aminopropyl silyl bonded phase (5 pmn Supelcosil LC-NH2, cat # 508338, from
Supelco; an Adsorbosphere NH 2, 3 pm column, cat # 287283, or Carbohydrate, 10 pm
column, cat # 35642 both from Alltech). The two Alltech columns differed in bead size,
and stability. The Alltech Carbohydrate column is more highly cross-linked than the
Alltech Adsobosphere column, increasing its stability to hydrolysis. Column dimensions
173
in all cases were 25 cm x 4.6 mm I. D. All three columns gave similar retention times for
monosaccharide standards. Columns were operated at room temperature using a degassed
mobile phase of 75/25 acetonitrile/water (v/v), at a flow rate of 1 ml/min. Attempts were
made to increase the relative amount of water in the mobile phase (to insure complete
dissolution of monosaccharides), but rapid loss in peak resolution was observed. Thus,
we found that 75% acetonitrile gave us the best combination of resolution, recovery and
sugar solubility. The mobile phase must be delivered in a capped solvent delivery system
to prevent evaporation of the acetonitrile, which will cause a rapid drift of the RI
detector.
The samples were dissolved in the mobile phase prior to injection. After sample
collection, the acetonitrile was removed by evaporating under nitrogen using a TurboVap
LV Evaporator (Zymark) at 50 OC. The remaining water was lyophilized, and a fraction
(5-10 %) of each sample was re-injected onto the amino column to insure that the
separation of was complete.
Retention times and reproducibility in peak areas were calibrated by sequentially
injecting larger sample amounts. Up to 200 jtg of each monosaccharide was injected
with 100% recovery off the column. As shown in Figure 4.4A, retention times for seven
consecutive injections of increasing sample size (5-100 jig of each sugar/inj) did not shift
more than ±0.5% for each sugar. The standard deviation for the calculated areas under
each of the individual monosaccharide peaks ranged from 5-12%. The calibrated
response factors differed for individual monosaccharides, with arabinose, mannose and
galactose showing lower responses than rhamnose, fucose, xylose and glucose.
174
i
Fucose and Xylose
250
2 00
i
Mannnose and Glucose
Rhamnos
Arabinose
150
/
,
Galactose
100
/Y~x.;,Y
----..............
04
i
i
2
8
46
175
10
12
14
250
= 0.9994 (Fucose + Xylose)
200 -
X
150
-
= 0.9985 (Arabinose)
R2= 0.9985 (Arabinose)
R2= 0.9988 (Rhamnose)
R2 = 0.997 (Mannose + Glucos
/g
A-
-
100 -
R2 = 0.9958 (Galactose)
50
0
0
2000
4000
6000
8000
10000
Area
12000
14000
16000
18000
20000
Figure 4.4 (A) Reproducibility of six consecutive injections of a monosaccharide mixture on the
amino column. Standard deviations (% Stud. Dev) for retention times (Ret. Times) and peak areas
are reported as a percentage. (B) Calibrated detector response for five monosaccharide fractions (R2
for each linear fit to the data are reported on the figure).
In the case of UF961, CRL, and P961, purified monosaccharide fractions were reinjected onto the ion exchange column to remove bleed from the amino column.
176
4.2.4
Sample Preparationfor Accelerator Mass Spectrometry (AMS)
After the final HPLC purification step (Figure 4.2), samples were lyophilized and
re-dissolved in 1-2 ml of Milli Q water. Using pre-combusted 13" NMR pipettes (Cat #
803A, Wilmad Glass Co,Buena, NJ), samples were transferred to the bottom of 8" Vycor
quartz tubes (9 mm O.D x7 mm I. D., Anderson Glass Co,Inc, Fitzwilliams, NH). These
quartz tubes, containing 300-500 mg of CuO (copper oxide wires, fine wires 4 mm x 0.5
mm (sieved), Elemental Microanalysis, Manchester, MA, (stored in a dessicator)), had
been combusted at 850C for 5 h prior to the addition of each sample. Care was taken to
avoid dripping the samples along the side of the quartz tubes in order to minimize losses
during the flame sealing step. The copper oxide was pre-weighed using an aluminum boat
and placed at the bottom of each combustion tube using a long-stemmed funnel. All
weighing equipment was sprayed with an Aero Duster® (Miller Stephson Chemical
Company, Danbury, CN) to remove any dust. The samples, in quartz tubes, were
lyophilized overnight, covered with pre combusted aluminum foil (450°C) and flame
sealed the following day.
Prior to flame sealing, quartz tubes containing the sample and CuO were attached
to a vacuum manifold and evacuated. Connections were first cleaned using the Aero
Duster®. The samples were gradually placed under vacuum using a needle valve and
evacuated for at least 1 h. Prior to sealing, the samples were closed off to the vacuum
pump and the bottom 2" of each tube was placed in a liquid nitrogen bath to avoid any
sample loss through volatilization. Samples were then sealed using a butane/oxygen
torch. All tubes were annealed in the mixed gas flame, followed by a butane flame, for
one minute after the sealing was complete. The pyrolitic decomposition of CuO, yielding
oxygen for the oxidation of sugar carbon to CO 2 , occurs at temperatures above 500 'C
177
(Boutton, 1991). Thus, sealed samples were combusted at 850C for five hours. In the
case of larger samples (>300 p.g), or samples likely to contain salts, ceramic blocks were
used to separate individual tubes in the muffle furnace to avoid damaging surrounding
tubes in case of an explosion. The copper that is reduced during the oxidation of organic
matter will also help strip halogens such as chloride. The total HMW DOM sample tubes
contained 100 mg of Ag (silver shot, Alfa Aesar, Ward Hill, MA), in addition to CuO, to
react with other sea salt contaminants.
The CO 2 generated during the combustion was quantified 2-3 days later. Sample
tubes were scored, placed in crackers, and connected to a vacuum manifold. Samples
were cracked individually and CO 2 was quantified by measuring the partial pressure of
the gas in a pre-calibrated volume at a known temperature. Prior to quantification, the
gas was passed over a dry ice/isopropanol slush bath to trap any water produced during
the combustion. The CO 2 was then trapped in a liquid nitrogen bath and any remaining
gas (i.e., gases with freezing points lower than -200 0 C) was removed by evacuating the
manifold. Quantified CO 2 samples were then transferred to 6 mm pyrex tubes (Anderson
Glass Co, Fitzwilliams, NH) by trapping the CO 2 in the bottom of the tubes with liquid
nitrogen, and flame sealed. If splits of gas were taken for 813C analysis, the samples were
allowed to equilibrate for 3 minutes to prevent any isotopic fractionation.
Samples with > 70 pg of carbon, (the seven blanks for CRL combined as one
sample, all the individual monosaccharide samples from CRL, and rhamnose, fucose and
xylose from MAB961) were submitted as quantified CO 2 to Lawrence Livermore
National Laboratory (Livermore, CA) and prepared into graphite and dated by Dr.
Michaele Kashgarian. Smaller samples (combined blanks from MAB961, combined
178
blanks from P961, individual monosaccharides from P961, and both the blank and the
combined monosaccharides (rhamnose , fucose and xylose) from P962), were prepared
into graphite according to the microscale AMS sample preparation protocol (Pearson et
al. 1998). In all cases CO2 was reduced to graphite on a 325 mesh spherical cobalt
catalyst, that was cleaned at 4000 C for 0.5 h under 0.7 atm H2. A carbon:cobalt ratio of
60 tg/ 1 mg was used, but no less than 1 mg of Co was used (thus smaller samples have a
lower carbon to cobalt ratio). The graphite was pressed into targets by NOSAMS, and
analyzed along with standards and process blanks. Two primary standards NBS Oxalic
Acid I (NIST-SRM-4990) and Oxalic Acid II (NIST-SRM-4990C) are used during all 14 C
measurements. The process blank material is a Johnson-Mathey 99.9999% pure graphite
powder.
All the radiocarbon data reported in this chapter were determined by AMS, and
are reported as fin: fraction modern, or fm*: fraction modem blank-corrected (MAB961
and P962). Fraction modem is a measurement of the deviation of the 14C/12 C ratio of a
sample from "modemn" values. Modem is defined as 95% of the radiocarbon
concentration (in 1950) of NBS Oxalic Acid I normalized to 813CvPDB = -19 %o(Olsson,
1970). AMS results are calculated using the internationally accepted modem value of
1.176±0.010x10 - 12 (Karlen et al. 1968) and a final ' 3 C correction is made to normalize
the sample fin to a 813 CVPDB value of-25%o.
Thus fin is calculated according to:
AsN
fn=-/
Aabs
A ( 1 225 +513C
s
(1)
1000
I 2(19+8 13C
0.95AO 1000
ox y
1000
179
where As is the sample activity, Aox is the activity of oxalic acid, X is the decay constant
for 14C (1/8267 yr'), (Stuiver and Polach, 1977). The quantity fm* is calculated as:
fm*
-
f
sample
fm
M
Mblank
blank Msample
M
(2)
sample
sample
blank
Msample is the carbon mass of the total sample, and Mblank is the carbon mass of the
corresponding blank. Masses are determined after combustion to CO2 . In order to
convert fin to A14C and radiocarbon age, the Libby half life of 5568 years is used (
=
1/8033 yr' ). Also the present time, y, is taken to be 1997. Thus:
A14C =
14
*xeA(1950-y) - 1)x 100
CAge = 8033 In(fmn)
(3)
(4)
A 813C value of -22%o was used for all the monosaccharides, HMW DOC
samples and blanks. Blank samples were too small to remove a split for 513C analysis,
thus a value of-22%o was assumed. This assumption may be incorrect, but the age
correction is approximately +16 years for every +1%o difference from the standards'
-25%o value, which in this case is only a small fraction of the radiocarbon age of the
blanks.
Radiocarbon values for the blanks were -770 %oprior to the final purification (Ag
column) (Table 4.3). Following the final purification steps, blanks for CRL were
-431.4±10.3 %o. Blanks for the ultrafiltration cartridges were -999.5 %o (Santschi et al.,
1995).
180
43
Results
A list of recoveries following each of the purification steps is reported in Table
4.2. All samples (with the exception of MAB966 and MAB967) were desalted using an
anion exchange resin. Between 20-40% of the total sample applied on to the column was
lost during desalting (Desalting I). Consistent with this removal of material following
desalting, the chromatograms showed a marked difference in the size of the peak eluting
in the dead-volume of the column, as shown in Figure 4.5 (CRL (20 mg ) sample).
Although good separation of the monosaccharides from the unhydrolyzed material is
achieved in both steps, desalting via anion exchange, reduces the area under the peak
eluting at 6 minutes in (B) by 90 %. Monosaccharide recoveries, following desalting,
were between 70±10%. Other investigators have also observed similar losses of
monosaccharides during desalting using a mixed ion exchange (40-50 %, (Borch and
Kirchman, 1997)). No systematic relationship between the amount of sample lost and the
total amount of carbon applied onto the anion exchange resin was observed.
Following desalting (some samples were desalted twice) and filtration, the fraction of
monosaccharide carbon recovered from total HMW DOM ranged from < 1%to 5%. Of
the total carbon recovered following the initial desalting step (Desalting I),between 4-5%
and 1.5 -2.5% of the carbon in the surface and deep water samples were recovered as
monosaccharides. Recoveries of monosaccharide carbon following the final separation
on the amino column were approximately 50% lower than silver column recoveries.
AO1 A. ALDI IMWL A (LUmNW*13.0)
mAU
100-4
90
70
80
60.i
40
In
.)
1
---1
A30
A.A
ICH
N.A(
1
6
A
1
12
1
16
1,8in,
ADC1A. ADC CANI.A (LHNmNH164.)
vAUJ
100~
Xylose, Mannose and Galactose
Glucose and Rhamnose
Fucose and
901
on
Arabinose
,i
Fcs
n
aol
8o
70-
50-
Figure 4.5 CRL sample after hydrolysis and neutralization, separated on the cation exchange
column using Ag as the counterion. The figure shows the sample before (A) and after (B) desalting
using an anion exchange resin.
182
iii N11ueA
uesallng
ii
GFIF Filtering
mg C (Sugars)
M9 %. IU
Ag
CSoumn
Sample
HMW DOM
Mg
HMW DOM
mg C
Desalting I
Desalting I
mg C (Total) mg C (Sugars)
CRL
217
58.6
36.3
MAB961
210.6
82
67.2
4.0$
MAB966* 61.6
17.9
14.3
0.7
MAB967
84.4
24.5
P961
145.8
37.9
30.3
P962
167
36.9
29.1
mg C (Sugars)
IIII
mg
UUEI
r rllhUa~AII5U9~I
rrU
C (Sugars)IUIII
mg
C
mg C (Sugars)
(Sugars)
mg C (Sugars)
0.48
0.7
0.5i
0.25
0.40
0.4#
10.8
15.9
234
SS971
* MAB966 was collected using a cation exchange column in the Pb form not the Ag form
# P962 was also treated with AgNO 3 to precipitate chloride.
$ Recoveries are the sum of a fraction collected after Desalting I and another collected after Desalting II.
& These samples had high blanks from indicating contamination.
0.2i&
0.2&
Table 4.1.2 The recovery of carbon during each of the different purification steps performed for this analysis. Desalting was performed on an anion
exchange column. Ag column and Amino column data, refer to recoveries during the HPLC procedure
183
The deep Pacific Ocean sample (P962) had the lowest carbon recovery. In
addition to being desalted twice and filtered, it was also mixed with 10 ml of a saturated
solution of AgNO 3 (dissolved in Milli Q water) to try and precipitate the chloride ions
prior to HPLC. The precipitate was removed by centrifuging the sample and pipetting
out the water (repeated 3-4 times). The AgC1 precipitation was performed because the
deep sample processed prior to P962 (SS971) stripped Ag ions from the cation exchange
column as AgC1.
In all cases, the low recovery of monosaccharides is likely due to the presence of
high concentrations of cations in these samples, which interfered with the separation and
generated a mixed ligand cation exchange column. Also, the use of the anion exchange
column in the C1- form for desalting the samples introduced large concentrations of
chloride ions to the cation exchange column, causing the removal of Ag ions as AgC1.
The relative distribution of monosaccharides and the amount of carbon in HMW
DOM accounted for by these monosaccharides was also determined by gas
chromatography (GC). The method and the results of the analysis for UF961, P961 and
P962, and UF967 are reported in Chapter 2. Briefly, samples were hydrolyzed with 2M
TFA as in Section 4.2.2, but were not desalted or filtered. Monosaccharides were
derivatized and analyzed as alditol acetates by GC. In all cases, monosaccharides were
shown to account for 15% of the carbon in surface water and between 4-6 % of the
carbon in deep water HMW DOM. Comparing this data to the yields in Table 4.2, it is
clear that the HPLC method underestimates the relative amount of monosaccharides in
HMW DOM compared to the GC analysis. For comparison, monosaccharide recoveries
in other studies range between 6-20% in surface waters and 1.6-3.5% in deep waters
(Borch and Kirchman, 1997; McCarthy et al., 1996; Skoog and Benner, 1997). In all
cases, quantitation of monosaccharides by GC (this study and McCarthy et al. (1996)
gave higher yields than HPLC (Borch and Kirchman, 1997; Skoog and Benner, 1997)).
For deep samples, carbon was added between the separation of monosaccharides
on the amino column and the final combustion step (Table 4.2). This resulted from the
contamination of the samples and blanks with bleed from the amino column (propyl
amine was observed in the waste fraction by 1HNMR spectroscopy). The combustion
yields for contaminated individual blanks (galactose blanks from MAB961 and P961)
and combined blanks (P962 and SS971) were between 34-53 jig and 90-104 gg carbon.
The surface ocean samples (MAB961, CRL, and P961) were subsequently re-injected
onto the Ag column to remove the bleed from the amino column. For these samples,
further losses of carbon were incurred between the amino column and the final
combustion of the sugars to CO 2. Following the final clean-up step, blanks for CRL were
5.7±1.18 jig of C, 6.72+0.13 jig for MAB961 and 4.91±0.82 jig for P961. Blanks for the
lyophilizing system were 2.2 jig of carbon, and combustion blanks were <1.0 jig of
carbon.
High-performance liquid chromatography of all blanks showed a peak eluting
between 17-18 minutes on the silver column which was 5%of the area of the total
monosaccharide fraction. This peak was also present in the blanks of each of the
different hydrolysis procedures that were tested, indicating that it did not result from
either a contaminant in the TFA or the ultrafiltration and diafiltration procedures. In the
case of samples (UF961, P961 and P962) that were filtered using a Whatman GF/F filter,
an additional peak was detected, eluting 13-13.5 minutes. Individual monosaccharide
185
blanks on the amino column showed a peak eluting at approximately 6.3 minutes
(between the rhamnose and the fucose + xylose peak).
Shown in Figure 4.6 A is a chromatogram of the CRL sample during the initial
separation of monosaccharides from unhydrolyzed material (Ag). All samples were
characterized by a large initial peak and a large baseline. This baseline was present in the
monosaccharide region of the chromatogram as well, and was thus collected with the
monosaccharide fraction. Figure 4.6B shows the re-injection of each of the collected
monosaccharide fractions (F , F2 and F3) onto the silver column for further purification.
Subsequent re-injection of the monosaccharide fractions (F , F2 and F3) shows the
presence of some of the baseline material following the initial collection. The second
injection of CRL onto the cation exchange column greatly increases the separation of the
monosaccharide fractions from the total unhydrolyzed material.
Following the Ag column separation (as shown in Figure 4.2), each dried
monosaccharide fraction was injected re-dissolved in 75/25 acetonitrile/water (v/v) and
injected onto the amino column. Figures 4.7 A through C show the separation of each of
the fractions (F1, F2 and F3) collected from the Ag column by reverse phase
chromatography (amino column). As shown in the Figure 4.7, F1, F2 and F3 separate
well into their individual monosaccharide components on the amino column allowing
clean separation individual monosaccharides.
Subsequent combustion of some blanks and samples (as above) showed
contamination from the amino column. Thus the remaining individual monosaccharide
fractions were re-injected and separated by cation exchange chromatography (Ag
column) (Figure 4.2). This separation is also shown in Figures 4.7 A though C.
186
; sl--
---*-------'---"--.' "--. -"~
~
~'
"~I~~L--~..-cY------------~Y~~ ~^~'""~"*
'~~ ------ ------ -~~~'~'I~"
M"~~"~ ---r"---- -----cl----------rr~--
I
A (LHACNO.D)
ADC1 A, ADC1 CHANNEL
rnau
oo
Fl
.
F2
F3
B
F2
F1
F3
75
70
651
soi
55 4
CRL (2"Ag
injection)
/
-
--- ----
-
4
B
in
12
14
16
13
Figure 4.6 HPLC of CRL on the Ag column, during the initial separation (A) and the second
separation (B)
-s
187
-
AUQ1 A, AUM1
ANNW. A (L3HIFfLO.0)
MAU
=s
F1
F1
200
CRL (Ag Column)
100-:
loo~
O
mAU200-
Rhamnose
Glucose
CRL (Amino Column)
S
~r
--
-IV1----
50
0-
~ ~6
~
AUA.lA ALA-lUrIAVM
~
~
~
A tLrWNIWAI.L)
mAU160-
SRhamnose
140
CRL (Ag Column)
\~
V
188
I
1'4
-
i
6 N
U __ M ~I LLLI nurme~
CRL (Ag Column)
i
I
-
189
I
ALIMi A. Al.]i
%.,mrlww. A kINWA.
mAU j
F3
661
I
f4-
CRL (Ag Column)
60-i
58
2
18
16
1
12
10
8
4
A=1 A, AU31 WANEL A (LWfNMNMl 1.U)
MAU
Fucose
CRL (Amino Column)
Aaio
Arabinose
4020-
-20
14
12
10
8
6
2
ALL1A.
AC CHANNL A (LNWLAB16FUD)
A (LHWALAB17.0)
ADCI A ADC1 CHANNEL
SADC A.
No
300I
/I
Arabin
250-
Fncose
i
I;
CRL (Ag Column)
1\
100k
50
.. 1
'
4
m
ji?
i
I
1
'I
1
A
In
12
1AR
,I_ .
Figure 4.7 Th separation of individual monosaccharides from unhydrolyzed HMW DOM using a
combination of cation exchange and reverse phase HPLC. (A) Separation of glucose and rhamnose;
(B) separation of mannose, xylose and galactose; (C) separation of fucose and arabinose.
190
4.3.1
Radiocarbon Analyses
Shown in Table 4.3 are the radiocarbon results for total HMW DOM and the
monosaccharide fraction of HMW DOM isolated from each sample. The Al 4C values for
the DIC and DOC in both the Atlantic and Pacific Ocean were obtained from Dr. James
Bauer (unpublished (Atlantic) and Bauer et al. 1998). The individual monosaccharides
isolated from the two Atlantic Ocean surface samples, CRL and MAB961, had modem
A14C values (between 0 and 116 %ofor CRL and between 49 and 92 90 for MAB961). In
both cases, A14C values of the individual monosaccharides and DIC (67.3 and 59 %o,
respectively) are in good agreement. The A14C values of the monosaccharides isolated
from CRL were also in good agreement with the A14C value of HMW DOM. The HMW
DOM fraction is considerably enriched in A14C at this site (95 6o) and, due to the
proximity of this station to shore, may reflect a terrestrial DOM source. The
monosaccharides isolated from MAB961 were enriched in A14C compared to the total
The HMW DOM and the total
HMW DOM fraction (-10 %o) by approximately 60 9/%o.
DOM at this site are also enriched in 14C compared to open ocean values (typically -200
9oo) and likely reflects the presence of some terrestrially derived DOM (from the
Delaware River and the Chesapeake Bay).
The results from the MAB profile (MAB966 (300 m) and MAB967 (750 m))
indicate that the sugars in HMW DOM at both these depths (-120 and -59 96o,
respectively) are younger than the total HMW DOM fraction (-375 and -255 %o,
respectively), but older than the DIC (-32 and -33 %orespectively). HMW DOM is
enriched in 14C compared to DOM at both these depths, indicating that HMW DOM
CRL
DIC
Total DOC
HMW DOC
HMW Sugar Blank
Rhamnose
Fucose
Xylose+Arabinose+Galactose
Glucose+Mannose
MAB961
Weight
A"4C
ErrorAM"C
67.3
112.64
83.85
88.4
132.8
95
-431.4
116.2
1.0
58.0
84.8
5.2
10.3
11
12.2
11.7
10
Weight
A7 4 C
Error MC
59
-32
-10.2
6.2
DIC
Total DOC
HMW DOC
HMW Sugar Blank
Rhamnose
Fucose
Xylose
68.8
110
79.5
49.4
71.9
92.3
17.9
11.4
14.6
MAB966
Weight
A 4C
Error A14 C
DIC
Total DOC
HMW DOC
Total Sugars
720
-32
-414
-375
-120.3
4.4
4.5
MAB967
Weight
A 4C
ErrorA"C
DIC
Total DOC
HMW DOC
Total Sugars
290
250
-33
-405
-254.7
-59
5.9
7.7
P961
Weight
A74C
Error A 7C
68
-275
-37.6
4.4
80.3
15.5
33.8
20.1
DIC
Total DOC
HMW DOC
HMW Sugar Blank
Rhamnose
Fucose
Xylose
Arabinose
Galactose
54
128.4
65.5
88
41.7
P962
Weight
A74C
ErrorAC
102.6
32
-234
-556
-351
-772.2
56.5
5.6
3.9
70.0
DIC
Total DOC
HMW DOC
HMW Sugar Blank
Rhamnose+Fucose+Xylose
Table 4.1.3 The A' 4C values of different carbon fractions at each of the sites examined in this
chapter. Refer to the text for descriptions of fm, fm* and A' 4C.
192
may be enriched in a young component. As stated earlier, these two samples were
collected after separation using a cation exchange column only, with no subsequent
purification of the individual monosaccharide fractions. Thus they are likely to contain
compounds other than monosaccharides.
The data for the Pacific Ocean profile are also given in Table 4.3 (P961 (surface)
and P962 (1600 m)). Both fucose and arabinose, isolated from the surface ocean HWM
DOM sample, have modem A14 C values (33 and 80 %o)which compare well to the
surface DIC (68 %o), and are enriched over the total DOM and the HMW DOM fraction
(-275 and -37 %o,respectively). In the case of HMW DOM isolated from the deep
ocean, the A14C value of the rhamnose + fucose+xylose, (56.5±70 $60) fraction was
enriched over the total DOM, HMW DOM and DIC (-556, -351 and -234
%/o)at
this
depth. The A14C value of these samples did agree well with the values for the individual
monosaccharides in the surface ocean and the surface DIC. Consistent with the Atlantic
Ocean data, the HMW DOM fraction is enriched in A14C compared to the total DOM,
indicating that the fraction with a modem A14C value (e.g. sugars) may be enriched in the
HWM DOM fraction.
In summary, all surface ocean monosaccharide samples showed modern A 4 C
values that compare well with the A14C value of DIC at each station. Monosaccharides
were also enriched in 14C compared to HMW DOM. In the deep Atlantic Ocean, the
monosaccharide fraction was younger than the HMW DOM but older than the DIC at
each depth. In the deep Pacific Ocean, on the other hand, the combined monosaccharide
fraction was much younger than the HMW DOM and the DIC (at 1600 m) and compares
193
well to the A14 C value of surface ocean DIC. For all samples HMW DOM is enriched in
A14 C compared to total DOM.
The similarity in the A'4 C values of the monosaccharides and DIC (modem
values), coupled with the fact that the contaminant in our samples is old (-772 prior to the
final silver column injection (contaminant from the amino column), and -432.4 after the
final purification and isolation), indicates that the data obtained by this method are robust
and not an artifact of the analysis.
The errors reported in Table 4.3 are largely due to small sample sizes. In most
cases (except P962) the largest error over the entire procedure is associated with the AMS
measurement. The sample weights and blanks were well quantified (by combustion) for
the surface samples so there was little error associated with this step. The errors in the
AMS measurement are usually associated with the counting: (1) either statistical errors
associated with the number of counts measured for each target or (2) errors associated
with the reproducibility of the of individual analyses for a given target. The deep Pacific
Ocean sample was combusted before the contamination from the amino column was
detected and as such had a blank of 100 jlg of C (A' 4C values of-770). In this case, the
largest error for the entire analysis was that associated with quantifying the total amount
of carbon contributed by the sugars to the sample. The weight of the sample, determined
by integrating HPLC chromatograms, was 32±5 jtg C, giving a A14C error of
approximately ±70 %o.
194
4.4
4.4.1
Discussion
Radiocarbon Results
Structural studies in Chapter 2 identified the presence of APS with a fixed
monosaccharide distribution in both the surface and the deep oceans. Up to 60% of the
surface water HMW DOC and 20% of the HMW DOC in the deep ocean is estimated to
be APS. This decrease in APS concentration with depth, coupled with the data presented
in Chapter 3 indicate that APS is produced during primary productivity in the surface
ocean. The cycle of APS in the ocean must thus be controlled by production in surface
waters, followed by transport to depth by one of the two mechanisms identified in section
4.1 (Figure 4.1), namely, the formation of deep water at high latitudes or the injection of
DOM into the deep sea by particle dissolution.
We can use the radiocarbon data compiled in this chapter to determine the
importance of these two mechanisms in controlling the cycle of APS. Shown in Figure
4.8 is a compilation of the Al 4C values in different fractions of carbon at Station M,
where the Pacific Ocean profile (P961 and P962) used in this study was collected.
The figure shows a depth profile of A'4C values for both DOC and DIC, and the A14C
values of suspended POC (at 1600 m) and sinking POC (600 m above the bottom)
between 1991 and 1992 (Bauer et al., 1998; Druffel et al., 1996). In addition, the figure
also shows the A14C values determined in this study for total HMW DOC and the HMW
sugar fraction (APS) in both the surface ocean and at a depth of 1600 m. Even though
the DOC is consistently 250 %oolder than the DIC, the depth profiles of both DIC and
DOC have a similar shape, indicating that the vertical distribution of these two fractions
is controlled by the same mechanism.
195
A14 C
-00 -500 -400 -300 -200 -100
0
O
0
100 200
500
1000
HMWSugs
1500
2000
DIC
IDOC
POCs
rDth(n)
2500
3000
3500
sink
4000
4500
*
HMW DOC
*
HMW Sugars
-POC
Figure 4.8 Radiocarbon values for different fractions of carbon in the North East Pacific (Station
M). POC., is the A14C value of suspended POC, and POCsik is the A14C value of sinking POC.
Depth profiles show DOC and DIC A14C values; filled circles are the A14C value of total HMW DOC
(surface and 1600 m), and filled diamonds are the AI4C value of the HMW Sugars (surface and 1600
m) DIC, POC and total DOC values were obtained from Druffel et al. 1996 and Bauer et al. 1998.
In both the surface and the deep ocean, HMW DOM is consistently younger (-59
compared to -275 %oin the surface oceans, and -351 versus -556 %oat 1600 m) than the
total DOM. The monosaccharide fraction of HMW DOM, and thus the APS fraction, is
196
in the deep ocean and between 30-80 %oin the surface
additionally enriched in 14C (56 6%o
ocean). Based on the modern A 4 C values of the monosaccharides at 1600 m (compared
with DIC and DOM), the mechanisms which transport DIC and total DOM to the deep
ocean must be independent of the mechanisms which transport APS to the deep ocean.
The A' 4C values of APS (reported as HMW Sugars) at both depths compare well with
each other and with the A14 C values of surface DIC, suspended POC, and sinking POC.
Based on the similarity in A14C values of APS and the different POC fractions, this data
suggests that APS is transported to depth primarily by the dissolution or hydrolysis of
POC, and not by the subduction of water masses at the poles.
Druffel et al. (1996) have also observed decreasing A14C values for suspended
POC with depth (at this station), indicating either the adsorption of old DOC on to
particles or the removal of a younger POC component with depth. If the latter case is
true then the average age of suspended POC would decrease as the younger POC
component is removed and the older component remains. The data for APS presented in
this chapter suggests that a new component is indeed stripped from the partciulate phase
into the dissolved phase, in the deep ocean. This stripping of new APS from the
particulate phase may serve to age the suspended POC component with depth.
Past studies on the radiocarbon ages of DOM have focused on the total DOM
fraction (Bauer et al., 1998; Druffel et al., 1992; Williams and Druffel, 1987) and have
found that A
-230%0.
4 C values
in the surface Atlantic and Pacific oceans range from -150%0 to
The A14C values of DOM isolated from the deep ocean range between -400%0/
(4000 radiocarbon years), in the Atlantic Ocean to -525960 (6000 radiocarbon years) in
the Pacific Ocean (Druffel et al., 1992; Williams and Druffel, 1987). This 2000 year
197
inter-basin difference in the radiocarbon age of bulk DOM is similar to the 1500 year
deep ocean mixing time, and results from the passive aging of the DOM during deep
water mass transport. However, the data collected in our study shows the presence of
modem or post-bomb radiocarbon in a specific fraction of DOM (namely APS) isolated
from the deep ocean indicating that within the larger cycle of total DOM, smaller, more
dynamic cycles exist. Our data is consistent with the recent work of Hansell and Carlson,
(1998) which shows variability in the concentration of DOC in the deep ocean (Figure
4.9). The variability in the DOC concentrations at low and middle latitudes is
approximately 1-2 pM, which agrees well with the predicted concentration of APS in the
deep ocean (20% of HMW DOM, and thus 2-3% of the total DOC) based on the data
presented in Chapter 2. Our deep ocean data also indicate that unlike total DOM, which
shows inter-basin differences in A 4 C values, APS, delivered by particle dissolution,
should have similar A' 4 C values in both oceans.
48.1
33.8
43.6
43
43.0
.41,7
41.
Figure 4.9 Variations in the concentration of total DOC (pM) over the whole ocean (after Hansell
and Carlson, 1998).
198
Several studies exist in the literature indicating the release of DOM into the deep
ocean by sinking particles. The diffusion of DOM from rapidly sinking fecal pellets
(Jumars et al., 1989) and other particles has been hypothesized as one mechanism for the
introduction of labile components from non-living POM to the deep ocean (Ittekot et al.,
1984). Dissolved organic matter may also be produced from particles by attached
bacteria, due to loose hydrolysis-uptake coupling (Azam et al., 1983; Cho and Azam,
1988). Investigations of the extracellular hydrolytic activity of attached bacteria show
that polymer hydrolysis rates exceed the carbon demand of attached bacteria, suggesting
that DOM is being released into the surrounding waters more rapidly than it is being
assimilated (Karner and Herndl, 1992; Smith et al., 1992). Studies have also shown
elevated growth rates of free bacteria in the presence of colonized marine snow (Hemdl,
1988), and a spatial and temporal coupling of deep sea bacterial biomass with surface
production (Nagata et al., 1998). This data suggests that bacterial transformation of POM
to DOM and ultimately DIC, occurs in the deep sea. Our data shows direct evidence for
14
the production of DOM from POM in the deep ocean. Furthermore, the modern A C
value of APS in the deep ocean suggests that this fraction is turning over on short time
scales. If APS is indeed 20 % of the HMW DOM in the deep ocean, as indicated by the
structural data in Chapter 2, this corresponds to a deep ocean reservoir of approximately
20 Gt C, which is likely to be an important substrate for deep sea bacteria.
Chemical characterization of transparent exopolymer particles (TEP), (which are
believed to be an important constituent of marine snow (Alldredge and Gotschalk, 1988;
Mopper et al., 1995), POC (Cowie and Hedges, 1984; Skoog and Benner, 1997) and
POM gels (Chin et al., 1998) all show the abundance of polysaccharides. Figure 4.10
199
shows the distribution of individual monosaccharides in TEP and POM compared to
seawater HMW DOM. It is clear that these organic carbon fractions contain the same
seven neutral monosaccharides. The relative distribution of monosaccharides in HMW
DOM and TEP are very similar. If glucose is removed from the distribution, all three
fractions show similar monosaccharide distributions. Thus, both possible sources and
mechanisms for the rapid delivery of APS into the deep ocean have been identified. The
contemporary age of APS is evidence for the occurrence of these processes.
The similarity in the A14 C values of the individual monosaccharides (30-100%0)
and their similarity to surface water DIC confirm previous findings that these
monosaccharides are present together in a single polysaccharide, namely APS which is
produced in surface waters during photosynthesis.
100%
-
-
90%
80%
aX
70%
IG
60%
jmF
50%
40%
30%
20% -
10%
0%
Average
TEP Avg
POC
Figure 4.10 A comparison of the distribution of monosaccharides in HMW DOM (Average),
Transparent Exopolymer Particles (Avg TEP) ( after Mopper et al., 1995), and sediment trap
material (POC) (after Cowie and Hedges, 1984). Monosaccharides are galactose (Ga), fucose (F),
mannose (M), glucose (G), rhamnose and arabinose (R+A) and xylose (X)
200
The monosaccharides separated from the MAB depth profile (-120%o (MAB966,
300 m) and -599oo (MAB967, 750 m) had A14 C values older than the DIC at each depth
(-320o and -33%o), contrary to the data from the Pacific Ocean sample (P962). However,
these MAB samples were only separated on the cation exchange column and showed
considerable contamination (up to 60%) from the base line. Unhydrolyzed HMW DOM
must be older than the monosaccharide fraction in order to maintain a mass balance of the
radiocarbon. Even a small fraction of this old material in the monosaccharide samples
could decrease their A 4 C values considerably.
4.4.2
Analytical Methods
The overall method described in this chapter realizes the goal of separating
individual monosaccharides from total hydrolyzed HMW DOM for radiocarbon analysis.
However, several problems were encountered when developing this method. The
cumulative result of these different problems was low monosaccharide recoveries at the
end of the analyses shown in Figure 4.2.
Several avenues were explored in an attempt to increase monosaccharide
recoveries, including a modification of the hydrolysis and desalting procedures discussed
in section 4.2.
Two other hydrolysis methods were compared before the TFA method was
chosen. One method used 4M HC1 hydrolysis at 100 OC for 4 hours (10 mg of total
sample/ 2ml acid). The method of Cowie and Hedges (1984) was also tested, where the
HMW DOM samples were hydrolyzed with 12 M H2SO 4 for 2 h at room temperature,
then diluted to 1.2 M and hydrolyzed for 3 hours at 100 0 C. Both the TFA and HC1
methods gave comparable yields (7-9% of total carbon in CRL HMW DOM), while the
201
H2 SO 4 method gave lower yields (3-5%). Time course studies on the 2M TFA hydrolysis
of CRL HMW DOM showed no increase in monosaccharide yields between 40 minutes
and 2 h. We also attempted to increase hydrolysis yields by performing consecutive 2 M
TFA hydrolyses on HMW DOM. Samples were first hydrolyzed, then neutralized with
ammonium hydroxide (NH40H), and hydrolyzed again. There were no detectable
increases in monosaccharide yields by either HPLC or 'HNMR.
Two different cation exchange columns were also tested for the HPLC separation.
In addition to the cation exchange column in the Ag form (described in section 4.2), a
column using Pb as the counter ion (Supelco (SupelcogelTM Pb, cat # 5-9343, Supelco,
Inc, Bellefonte, PA) (referred to hereafter as the Pb column) was also tested. Both
columns were sensitive to the cations present in the samples, and while the Pb column
gave better separation between monosaccharides, it was also more sensitive to salt
concentrations in our samples and deteriorated quickly. Metal ions (Na+ , Fe 3+, Ca
+)
and
H+, could replace both the Pb and Ag ions, resulting in a mixed ligand-exchange column.
This changes the separation characteristics of the column and may also act to swell or
shrink the resin. Two methods of desalting (cation and anion exchange) were tested to
determine if the column lifetime and separation characteristics of each sample could be
enhanced. Monosaccharide recoveries, for a CRL sample, spiked with 0.5 mg of a
monosaccharide standard, following hydrolysis, neutralizing and lyophilization were
100± 10%, with no detectable changes in the relative distribution of the monosaccharides.
During desalting, monosaccharide recoveries were between 7010% for the anion
exchange resin, and 6-7.5% for the cation exchange resin. Due to the high
202
monosaccharide losses to the cation exchange column, samples were not desalted with
this resin.
While desalting with the anion exchange column enhanced the separation of
monosaccharides from the unhydrolyzed material, this method introduced high
concentrations of CI ions into the sample. The presence of C1- ions in the sample
interfered with the separation of monosaccharides in the Ag column over time by
stripping the Ag from the column as AgC1. In the future, samples should be desalted
using an anion exchange column in the hydroxide form.
The low monosaccharide recoveries obtained by this procedure were of concern
primarily due the possibility of isotopic fractionation associated with the loss of this
material. However, no systematic removal of any particular monosaccharide was
observed either during Ag column separation or amino column separation. Following the
Ag column separation, the relative distribution of monosaccharides in the samples was
similar to the relative distribution of monosaccharides determined by GC. Following the
amino column separation, however, changes in the relative distribution of
monosaccharides compared to the distribution determined by GC were detected. Figure
4.11 shows the relative distribution of monosaccharides in the three surface samples
examined in this study. While the relative distribution of monosaccharides determined by
GC and HPLC are different, there is no systematic removal of any particular
monosaccharide.
203
50
=
o
45J
DR
40
N F+X
35
MA
M+G
30
OGa
*:.
020
15
0
Figure 4.11 The relative distribution of monosaccharides in HMW DOM injected onto the amino
column compared to the relative distribution of monosaccharides in HMW DOM determined by GC
(AA).
Furthermore, the comparable ages of the individual monosaccharides and their
similarity to the A' 4 C value of DIC at each site indicate that there is no isotopic
fractionation associated with the losses of monosaccharides during the analysis.
4.5
Conclusions
This chapter describes a robust method for the isolation of individual
monosaccharides from hydrolyzed DOM for radiocarbon analysis. The samples
processed in this study were not collected specifically for this purpose and due to
extensive method development, showed low monosaccharide recoveries. However, none
of the steps in this analysis systematically fractionates the monosaccharide distribution.
Thus a representative sample of all the monosaccharides can be isolated.
204
The similarity in the A' 4C values of individual monosaccharides is consistent with
the structural data that showed a constant relative distribution of monosaccharides,
implying that these monosaccharides are present together within a single polysaccharide.
Furthermore, the similarity between the A14C values of surface DIC and the
monosaccharides isolated from HMW DOM indicates biosynthesis of these
polysaccharides from DIC during photosynthesis consistent with the data presented in
Chapter 3. The deep ocean data from the Pacific shows for the first time the presence of
bomb radiocarbon in DOM isolated from the deep ocean (in the form of APS). The
14C
enrichment of APS relative to DIC (at 1600 m) and the similarity in the A14C values of
APS and POM suggest that APS is injected into deep ocean by settling particles. The
contemporary age of the monosaccharides in the deep ocean further indicates rapid
cycling of APS, highlighting its importance as a substrate for deep sea bacteria. Finally,
the concentration of APS in the deep ocean is estimated to be 1-2 tpM (based on the
alditol acetate yields), on the same order of magnitude as the changes in the concentration
of DOC in the deep ocean observed by Hansell and Carlson (1998).
205
4.6
References
Alldredge A. L. and Gotschalk C. (1988) In situ settling behavior of marine snow.
Limnology and Oceanography33(3), 339-351.
Azam F., Fenchel T., Field J. G., Gray J. S., Meyer-Reil L. A., and Thingstad F. (1983)
The ecological role of water-column microbes in the sea. MarineEcology
ProgressSeries 10, 257-263.
Bauer J. E., Druffel E. R. M., Williams P. M., Wolgast D. M., and Griffin S. (1998)
Temporal variability in dissolved organic carbon and radiocarbon in the eastern
North Pacific Ocean. Journalof GeophysicalResearch 103(C2), 2867-2881.
Borch N. H. and Kirchman D. L. (1997) Concentration and composition of dissolved
combined neutral sugars (polysaccharides) in seawater determined by HPLCPAD. Marine Chemistry 57, 85-95.
Boutton T. (1991) Stable carbon isotope ratio of natural materials: 1. Sample
preparation and mass spectrometric analysis. In Carbon Isotope Techniques (ed.
D. C. Coleman and B. Fry), pp. 155-170.
Broecker W. M. and Peng T. (1982) Tracers in The Sea. Lamont-Doherty Geological
Observatory.
206
Chin W.-C., Orellana M. V., and Verdugo P. (1998) Spontaneous assembly of marine
dissolved organic matter into polymer gels. Nature 391, 568-572.
Cho B. C. and Azam F. (1988) Major role of bacteria in biogeochemical fluxes in the
ocean's interior. Nature 332, 441-443.
Cowie G. L. and Hedges J. I. (1984) Determination of neutral sugars in plankton,
sediments, and wood, by capillary gas chromatography of equilibrated isomeric
mixtures. Analytical Chemistry 56, 497-504.
Druffel E. R. M., Bauer J. E., Williams P. M., Griffin W., and Wolgast D. (1996)
Seasonal variability of particulate organic radiocarbon in the northeast Pacific
Ocean. Journalof GeophysicalResearch 101(C9), 20543-20552.
Druffel E. R. M., Williams P. M., Bauer J. E., and Ertel J. R. (1992) Cycling of dissolved
and particulate organic matter in the open ocean. Journalof Geophysical
Research 97(10), 15639-15659.
Hansell D. A. and Carlson C. A. (1998) Deep-ocean gradients in the concentration of
dissolved organic carbon. Nature 395, 263-266.
207
Herndl G. J. (1988) Ecology of amorphous aggregations (marine snow) in the Northern
Adriatic Sea. II. Microbial density and activity in marine snow and its
implications to overall pelagic processes. MarineEcology ProgressSeries 48,
265-275.
Ittekot V., Deuser W. G., and Degens E. T. (1984) Seasonality in the fluxes of sugars,
amino acids, and amino sugars to the deep ocean: Sargasso Sea. Deep Sea
Research 31(9), 1057-1069.
Jumars P. A., Penry D. L., Baross J. A., Perry M. J., and Frost B. W. (1989) Closing the
microbial loop: Dissolved carbon pathways to heterotrophic bacteria from
incomplete ingestion, digestion and adsorption in animals. Deep Sea Research
PartA 36, 483-495.
Karner M. and Herndl G. J. (1992) Extracellular enzymatic activity and secondary
production in free-living and marine snow associated bacteria. Marine Biology
113, 341-347.
McCarthy M., Hedges J. I., and Benner R. (1996) Major biochemical composition of
dissolved high molecular weight organic matter in seawater. Marine Chemistry
55, 282-297.
208
Mopper K., Zhou J., Sri Ramana K., Passow U., Dam H. G., and Drapeau D. T. (1995)
The role of surface active carbohydrates in the flocculation of a diatom bloom in a
mesocosm. Deep Sea Research II 42(1), 47-73.
Nagata T., Fukuda H., Fukuda R., and Koike I. (1998) Regional variability of
bacterioplankton biomass in the deep ocean and its implications for oceanic
carbon cycling. American Society ofLimnology and Oceanography,Meeting
Abstracts San Diego, 1998.
Santschi P. H., Guo L., Baskaran M., Trumbore S., Southon J., Bianchi T. S., Honeyman
B., and Cifuentes L. (1995) Isotopic evidence for the contemporary origin of high
molecular weight organic matter in oceanic environments. Geochimica et
CosmochimicaActa 59(3), 625-631.
Skoog A. and Benner R. (1997) Aldoses in various size fractions of marine organic
matter: implications for carbon cycling. Limnology and Oceanography42, 18031813.
Smith D. C., Simon M., Alldredge A. L., and Azam F. (1992) Intense hydrolytic enzyme
activities on marine aggregates and implications for rapid particle dissolution.
Nature 359, 139-142.
Stuiver M. and Polach H. A. (1977) Discussion: Reporting of 14 C data. Radiocarbon 19,
355-363.
209
Williams P. M. and Druffel E. R. M. (1987) Radiocarbon in dissolved organic matter in
the central North Pacific Ocean. Nature 330, 246-248.
210
5.
Conclusions
5.1
General Conclusions
Despite its large size and importance to the global carbon cycle, little is known
about the composition and cycling of dissolved organic matter (DOM) in the ocean. Our
understanding of the oceanic cycle of DOM has been impeded by the inability to isolate
sufficient quantities of DOM, without chemical fractionation, for the more sophisticated
analytical techniques such as 1HNMR spectroscopy. When this thesis work was begun,
only 5-10% of the total DOM had been isolated as humic substances by adsorption on to
hydrophobic resins (Gagosian and Stuermer, 1977). As discussed in the Chapter 1 the
extensive chemical characterization of these XAD isolates coupled with the old average
age of DOM determined by radiocarbon analyses (Williams and Druffel, 1987), led to the
hypothesis that DOM accumulating in the ocean was formed by the geopolymerization of
simple biochemicals.
Due to the advent of tangential flow ultrafiltration, it has been possible isolate
DOM based on size rather than chemical characteristics. Using filters with a nominal
pore size of 1 nm, it is possible to consistently isolate 20-30% of total DOM (high
molecular weight (HMW)) by ultrafiltration (Benner et al., 1992; Carlson and Mayer,
1985). This has allowed us and other investigators (Aluwihare et al., 1997; Benner et al.,
1992; McCarthy et al., 1996; McCarthy et al., 1998) to chemically characterize a more
representative fraction of DOM. The work presented in this thesis was undertaken with
the recognition that some knowledge of the molecular-level composition of DOM was
required in order to gain a better understanding of its cycle in the ocean.
211
Presented in Chapter 2 are the results from the chemical characterization of HMW
DOM isolated from several different sites including the Pacific Ocean, the Middle
Atlantic Bight, and the Eastern North Atlantic. Chapter 3 describes the isolation and
chemical characterization of HMW DOM from the cultures of three species of
phytoplankton: Thalassiosiraweissflogii, Emilianiahuxleyi and Phaeocystis. Chapter 4
addresses the use of radiocarbon analyses to determine the mechanisms that control the
vertical distribution of HMW DOM in the ocean. The results of this analysis are as
follows:
(1) Fifty to sixty percent of the HMWDOM isolatedfrom the surface ocean and 20-30%
of the HMWDOM isolatedfrom the deep ocean consist of chemically related,
identifiable biochemicals, at all the sites examined.
Chemical characterization using 1HNMR, '3 CNMR, 'SNNMR, monosaccharide
analysis, lipid analysis and linkage analysis, showed the presence of compounds
with a similar chemical composition at all the sites examined. These compounds
were termed acylated ploysaccharides (APS), and are composed of a similar ratio
of carbohydrate: acetate: lipid carbon (80±4:10±2:9±4). APS contains 7 neutral
monosaccharides in a fixed ratio. Rhamnose, fucose, xylose, mannose, and
glucose all show similar relative abundance (14±1%), with arabinose being
slightly depleted (6%) and galactose being slightly enriched (21%). Furthermore,
the 15NNMR data in Chapter 2 in combination with the C/N ratios in HMW DOM
suggests that APS also contain N-acetylated amino sugars, which constitute
approximately 50% of the carbohydrate fraction. Contrary to earlier beliefs that
212
DOM was uncharacterizable due to its polycondendensate nature, this study
indicates that the high resolution analytical methods available today can be used
successfully to determine the structure of a large fraction of the DOM isolated
from the ocean.
(2) Two of the three culturedphytoplankton species examined in this study, Thalassiosira
weissflogii and Emilianiahuxleyi, producedHMW DOM with similar chemical
characteristicsto HMWDOM isolatedfrom seawater.APSproduced in these
cultures were degradedmore slowly than otherpolysaccharidespresentin the
exudate at the beginningof the incubation.
The work in Chapter 3 showed that both Thalassiosiraweissflogii and Emiliania
huxleyi produced compounds that are closely related to APS isolated from
seawater. The Phaeocystis species examined in this study did not produce
detectable amounts of APS. Furthermore, degradation of the T. weissflogii
exudate showed that the component resembling APS was degraded more slowly
than other polysaccharides in the exudate suggesting an inherent resistance to
bacterial degradation. While the total amount of carbon associated with APS in
each of these cultures is likely to vary depending on growth conditions, this
particular study implies that APS may comprise 2-3 % of the total organic carbon
produced by these algae. Contrary to the geopolymerization model, this study
demonstrated that DOM accumulating in seawater might be directly produced by
biosynthesis.
(3) The A' 4 C values of the individualmonosaccharidesconfirm that the carbohydrate
fraction of HMWDOM (APS) is produced duringphotosynthesis in the surface
213
ocean. Furthermore,the APSfraction in the deep ocean is of contemporary origin
also, indicatingdelivery ofAPS to the deep ocean on time scales shorterthan the
mixing time of the ocean.
Chapter 4 describes a robust method for the isolation of the
monosaccharides that constitute APS from total HMW DOM for radiocarbon
analysis. The data show that the individual monosaccharides from each sample
have similar A 4 C values (50±30 %o)indicating that they were formed recently in
surface waters and are likely present together in a single polymer. These values
also compare well to the A14C of DIC at these same sites confirming that the APS
fraction is produced during photosynthesis. The A 4C values of APS in the deep
ocean are also similar to surface water DIC values indicating rapid transport of
APS to depth. Thus we hypothesize that APS produced in surface waters are
delivered to the deep ocean by settling and subsequent hydrolysis and dissolution
of particulate organic matter raining through the water column. This data implies
that APS in the deep ocean is turning over on annual to decadal time scales and is
the first evidence for the presence of bomb radiocarbon in deep ocean DOM.
Figure 5.1 shows a compilation of the data presented above into a simple model
for the cycling of APS in the ocean. This work shows that contrary to the accepted dogma
of geopolymerization up to 20% of the DOM accumulating in the ocean on annual to
decadal time scales is produced directly by biosynthesis. Furthermore, the radiocarbon
results suggest that while a major fraction of DOM is passively distributed throughout the
214
ocean via ocean circulation, particular fractions of DOM have more dynamic cycles and
may be injected into the deep ocean by particle dissolution.
CO2aun
10 GT
(4C = 50%.)
15 GT
(14C = 50%.)
Figure 5.1 Cycling of APS in the Ocean.
In order to better understand the rates associated with each of the arrows shown in
the Figure 5.1 a more comprehensive model (as in Figure 5.2) for the cycling of APS in
the surface ocean needs to be constructed.
215
Dissolved
Particulate
I
I
I
SPAPs =
k APS(Settling) [APS]
k(APS)degradation
Figure 5.2 Th cycling of APS in the Surface Ocean. See text for an explanation of the terms.
Using Figure 5.2, the time dependent change in.the inventory of APS in both the
surface dissolved phase and the surface particulate phase may be expressed as follows:
a[As]dsu
d
f
a[APS]i.r,
=p p
Pd As - k
= PPAPs
[APS]d -JAs
MS
(dgradation) [APS
+JdAPS
at
where [APS]dsuf and [APS]Pus
]d
_d
APS
0
=O
-k s(setfng)[APS]' = 0
-,s(Setling)
are the inventories of dissolved and particulate APS,
respectively, in the surface ocean (in Gt). pdAs and PPAPs are the production rates of
dissolved and particulate APS, respectively (in Gt C/yr). kAPS(degradation) and kAPs(setuing) are
the rate constants for the biodegradation of the dissolved APS and the settling of
particulate APS, respectively. S'APS is the sinking flux of POC, which is equal to the
product of kAPS(settling) and [APS]Psur . Combining the two equations we have
PdAPS + PPA s = kAPs(settling) [APS r + kAps (deg radation) [APS 1
216
The production rate (pdAPS) of dissolved APS can be estimated from the culture
data presented in Chapter 3. Dissolved APS concentrations in the cultures were
approximately 2% of the total POC plus DOC. Extending these values to the field and
using a value of 50 Gt C/yr for annual primary productivity in the ocean, the annual
production rate of APS (from the culture data) is approximately 1Gt C/yr. The
production rate of particulate APS (PPAps) can be estimated using the carbohydrate
content of phytoplankton. Dissolved APS contains rhamnose, fucose, xylose, mannose,
glucose and galactose in roughly equal abundance with arabinose being approximately
half as abundant and the other sugars (Chapter 2). Assuming that particulate APS also
contains these same ratios of monosaccharides and that all the rhamnose in the
phytoplankton cell is present only in APS, the rhamnose content of phytoplankton cells
can be used to estimate the total amount of particulate (or phytoplankton associated)
APS. Using the values for total POC and the percentage of rhamnose associated with this
POC in a laboratory culture of Skeletonema Costatum (Biersmith and Benner, 1998), it is
possible to estimate that 1.9-3 % of the total POC may be present as APS. The
biodegradation rate constant (kAPS(degradation)) can be estimated from the T. weissflogii
degradation experiment discussed in Chapter 3. Since the bacterial cell numbers in the
degradation experiment were higher than those observed in the field, the rate constant,
0.042/day (or 15.33/yr) calculated from the culture experiment is likely to be up to 2
orders of magnitude lower in the field. Thus kAPS(degadaon) is taken to be approximately
0.15 /yr. An average POC settling rate constant of 0.05/day (or 18.25/yr), was assumed
based on the data of Gustafsson et al. (1997). Sinking POC inventories are estimated to
be approximately 0.24-1.3 Gt C based on the data of Martin et al. (1997), and of this APS
217
is expected to be between 1.9-3% (as above for PPAps) . Each of the numbers used in this
model are given below in Table 6.1
Term
Value
[APS]dsurf
10 Gt C
[APS]Psurf
0.04 Gt C
pdAPS
1 Gt C/yr
PPAP
1.5 Gt C/yr
kAPS(degradation)
0.15 /yr
kAPS(setling)
18.25/yr
Table 5.1 Numerical values for each of the terms shown in Figure 5.2. See text for an explanation of
each of the terms and the sources of each of the reported values.
Substituting these values into
PdAPS + Pas
= kAPSttlg) [APS]
+k
eg~aon),[APS]d
we get
1 Gt C/yr + 1.5 Gt C/yr = 18.33/yr (0.04 Gt) + 0.15/yr (10 Gt)
These values show surprisingly good agreement indicating that the estimates from the
culture data are reasonable. However, many assumptions went into the generating these
numbers and in order to be certain of this data more work needs to be done in order to
218
establish robust production rates for both particulate and dissolved APS and
biodegradation rates for dissolved APS in culture.
In the deep ocean the only source of APS is the degradation of POC. Thus at
steady state, both the production of dissolved APS from particles and the degradation of
dissolved APS must be equal and can be expressed as :
kAPS(Setding) [APS]p
= kAPS2(degradaion) [APS]dep
where k APS2(degradation) is the rate constant for the degradation of dissolved APS in the
deep ocean and [APS]dep is the inventory of APS in the deep ocean (from Figure 5.1).
Substituting the appropriate values in the equation we can estimate that the degradation
rate of constant of APS in the deep ocean is approximately 0.05/yr.
5.2
Future Research Directions
Structural identification is likely to be the most powerful tool in understanding the
different processes that affect the concentration, distribution and residence time of
different fractions of DOM. Microbial degradation rates of DOM, the metal chelating
ability of DOM, the light absorbing properties and so on, are all controlled by the
chemical structure of DOM. Thus, a comprehensive understanding of any of these
processes requires an understanding of the molecular level composition of DOM. While
this study has made an initial attempt to do this, much work remains to be done, and
many avenues remain unexplored.
219
For example, 70-80% of total DOM (low molecular weight) remains
uncharacterized. If we are to understand the cycle of DOM, a concerted effort must be
made to design a method for the isolation of low molecular weight DOM from seawater.
Some similarities between the HMW and LMW fraction of DOM have been observed,
the C/N ratio of total DOM is similar to HMW DOM, between 15-22. Investigators have
also looked at the monosaccharide composition of the LMW fraction (Skoog and Benner
1997). However, little compositional information has been gleaned from these studies,
thus much work needs to be done in this area if we are to understand the composition of
LMW DOM.
The major challenge that remains from the work done in Chapter 2 is the
development of a method to isolate APS from the total HWM DOM fraction. Once this
is done, rigorous quantification of the individual components of APS may be performed.
The most promising method to do this may still be methylation of the polysaccharide.
Perfecting the methylation reaction will also aid in the linkage analysis, which is essential
for completely characterizing the polymer. Other possible methods for the isolation of
the polysaccharide fraction are gel electrophoresis, ethanol precipitation and affinity
chromatography. The source of lipid resonances in the 1HNMR needs to be identified,
perhaps by repeating the BBr 3 reaction. In order to establish the presence of N-acetylated
sugars (which are an important fraction of the HWM DOM as suggested by the data in
Chapter 2), a method to hydrolyze these polymers needs to be developed. Isolation of
APS from the total HMW DOM fraction may aid in this hydrolysis. Finally, this chapter
briefly addressed the possible presence ofunhydrolyzable proteins as the major fraction
of nitrogen in the deep ocean, and thus the major fraction of nitrogen in the entire ocean.
220
Nitrogen-15 NMR should be pursued as the major tool to characterize this fraction. In
order to establish the presence of proteins a similar hydrolysis experiment as was carried
out for surface water HMW DOM should be repeated for a deep ocean sample. Again,
gel electrophoresis may be a good tool for the final isolation of the protein fraction.
Many avenues remain to be studied in Chapter 3. From a physiological point of
view, variations in the production of APS according to changes in growth conditions may
offer some insight into the function of these compounds. It would also be interesting to
test if other Phaeocystis species can produce APS, particularly in light of the differences
in the quality of the DOM observed in regimes dominated by this species. While the
bacterial degradation of the exudates was extensively chemically characterized, much
needs to be done in determining the dynamics of the bacterial population during the
degradation of DOM. Not only should respiration, bacterial production and bacterial
numbers be monitored, but changes in species composition should also be monitored and
correlated with the removal of different fractions of DOM. This work can then be used to
correlate the vertical and horizontal variations in bacterial population with the chemical
quality of the DOM available for degradation in the ocean.
The radiocarbon method presented in Chapter 4 is a powerful tool for elucidating
the cycle of DOM in the ocean. However, the method needs some refinement. Some of
the carbon losses may be compensated for by collecting bigger samples, and avoiding
diafiltration. Furthermore, better desalting protocols should be developed in order to
preserve the life of the HPLC column and to avoid losing material to the column. In
order to conclusively establish the hypothesis that APS are released into the deep ocean
from POM (either due to hydrolysis or dissolution), depth profiles of the A' 4 C values in
221
APS isolated from both the Pacific and the Atlantic Ocean need to be constructed.
Further evidence for the release of APS from POC could be obtained by correlating small
variations in APS concentration with variations in particle flux. In order to do this
however, robust methods for the quantification of APS need to be developed. If the
concentration of APS is conclusively established, this can be used in conjunction with the
A' 4C values of HMW DOM and APS to determine the average age of the old component
of DOM present at each site. It would also be interesting to determine what fraction of
the DOM utilized by free living deep sea bacteria is derived from old refractory DOM
and young, newly synthesized DOM. An alternative explanation for the presence of
modem APS in the deep ocean is the biosynthesis of APS by deep sea bacteria.
It is
thus important to determine if deep sea bacteria produce sufficient quantities of DOM to
maintain the deep sea inventory of APS. Furthermore, exudates of bacteria should be
analyzed as in Chapter 3 to determine is bacteria are an additional source of APS.
222
53 References for Chapter 5
Aluwihare L. I., Repeta D. J., and Chen R. F. (1997) A major biopolymeric component to
dissolved organic carbon in surface seawater. Nature 387, 166-169.
Benner R., Pakulski J. D., McCarthy M., Hedges J. I., and Hatcher P. G. (1992) Bulk
chemical characteristics of dissolved organic matter in the ocean. Science 255,
1561-1564.
Biersmith A. and Benner R. (1998) Carbohydrates in phytoplankton and freshly produced
dissolved organic matter. Marine Chemistry 63, 131-144.
Carlson D. J. and Mayer L. M. (1985) Molecular weight distribution of dissolved organic
materials in seawater as determined by ultrafiltration : A re-examination. Marine
Chemistry 16, 155-171.
Gagosian R. B. and Stuermer D. H. (1977) The cycling of biogenic compounds and their
diagenetically transformed products in seawater. Marine Chemistry 5, 605-632.
223
Gustafsson, O., Gschwend, P. M., and Buesseler, K. 0. (1997) Settling removal rates of
PCBs into the northwestern Atlantic derived from 138U-134Th disequilibria.
Environmental Science and Technology 31, 3544-4550.
Martin, J. H, Knauer, G. A., Karl, D. M. and Broenkow, W. W. (1997) VERTEX: carbon
cycling in the northeast Pacific. Deep-Sea Research 34(2), 267-285.
McCarthy M. D., Pratum T., Hedges J. I., and Benner R. (1998) Chemical composition of
dissolved organic nitrogen in the ocean. Nature 390, 152-153.
McCarthy M. D., Pratum T., Hedges J. I., and Benner R. (1998) Chemical composition of
dissolved organic nitrogen in the ocean. Nature 390, 152-153.
Skoog, A. and Benner, R. (1997) Aldoses in various size fractions of marine organic
matter: implications for carbon cycling. Limnology and Oceanography42: 18031813.
Williams P. M. and Druffel E. R. M. (1987) Radiocarbon in dissolved organic matter in
the central North Pacific Ocean. Nature 330, 246-248.
224